Investigating Research Integrity

Proceedings of the First ORI

Research Conference on Research Integrity
ii
 Investigating Research Integrity
Proceedings of the
First ORI Research Conference on Research Integrity

editors

Nicholas H. Steneck, Ph.D.

Mary D. Scheetz, Ph.D.

[HHS Logo]

2002

iii
publication information page

iv
 Preface

Over the past twenty years, a consensus has developed that integrity is vitally important to the health
of federally funded research and that the key stakeholders–individual scientists, research institutions,
associations and societies, government sponsors, and the general public–all play important roles in
fostering research integrity. However, there is little consensus about the importance of and a lack of
empirical scientific evidence on specific problems than can and do undermine integrity in research.
Even those of us who are experienced in research integrity issues have in the past based too much of
our thinking on personal experience, personal and philosophical biases, individual case exposes, and
the public, political, and media response thereto. Accordingly, to advance to the next level in our
understanding, it is time for new approaches to the study and discussion of research integrity.
Since its establishment in 1992, the Office of Research Integrity (ORI) has conducted a number
of studies on research misconduct and research integrity, some of which are ongoing. The goal of
these studies has been to develop a knowledge base for addressing important research integrity issues,
including: the impact of misconduct allegations on exonerated scientists, the experience of
whistleblowers in the aftermath of making allegations, the content of research guidelines adopted by
medical schools, and the the incidence of research misconduct. Over time, it became apparent to ORI
that a more comprehensive, coordinated effort in collaboration with extramural research scholars was
needed to develop baseline knowledge for understanding research integrity issues. This recognition
led to the development of the first Research Conference on Research Integrity in November 2000 and
the revised papers published in this volume. ORI has also begun, with support from the National
Institute of Neurological Disorders and Stroke, a related “Research Program on Research Integrity.”
In the background report that begins this volume, Assessing the Integrity of Publicly Funded
Research, Dr. Nicholas Steneck (ORI’s consultant for the November 2000 conference and the related
research program) has summarized the state of the empirical research on research integrity. This
report provided important background information for participants at ORI’s Research Conference on
Research Integrity and for scholars and others in the research community generally.
The research conference background report and the conference papers published in this volume
will hopefully provide an important catalyst for identifying important problems and for improving
our understanding of research integrity issues. Although research integrity has been a high profile
topic for some twenty years and some important preliminary studies have been conducted, the
publications in this volume, while contributing valuable information, make clear how little we really
know about many key issues, such as: how often research misconduct occurs, what situations tend to
encourage or prevent it, how human subjects are best protected, how often conflicts of interest occur
in research and how they affect the integrity of the research, how common questionable research
practices are and what harm they cause to the research process, how students and research trainees
learn the ethics of science, and what career pressures or other factors influence their ability and desire
to follow the most honorable scientific practices.
These unanswered questions provide a significant opportunity for the Public Health Service and
the research community to build a knowledge base for examining research integrity through further
research. Research will permit us to understand in a more thorough and genuine way the influence

v
that research integrity issues have on the careers of scientists, the operation of research laboratories,
the generation of accurate and useful research results, and the confidence of the public and political
community in the research enterprise. It will also provide a science base for making important
decisions—by government, by research institutions, by the community of scientists, and ultimately by
the general public—in response to future research integrity issues and concerns that will inevitably
arise.

Chris B. Pascal, J.D., Director

Office of Research Integrity

vi
 Introduction
Researchers and research institutions are universally committed to maintaining high standards for
integrity in research. Precisely what this commitment entails, however, and whether it is being
fulfilled are questions that have not been subject to rigorous critical investigation. What is “research
integrity”? Can it be assessed? Do current research practices meet the high standards individuals and
institutions say they embrace? How are standards for best practices transmitted? Are current
approaches to fostering integrity appropriate and effective? Are all segments of the research
community appropriately contributing to the promotion of high standards for integrity in research?
Many individuals have provided answer to these questions, based on personal experience and
anecdotal evidence. Few scholarly studies have been undertaken to confirm or refute what is
commonly believed to be true about research integrity but is seldom demonstrated.
The papers published in this volume were originally presented at the first ORI Research
Conference on Research Integrity in Bethesda, Maryland, on November 19-20, 2000, and
subsequently reviewed and edited for publication. Abstracts for other papers and posters presented at
the conference but not published in this volume can be accessed at http://ori.dhhs.gov. Together, this
work represents the first comprehensive effort by a group of scholars to take a broad but critical look
at evidence underlying our assumptions about integrity in publicly funded research.
The organization of the Proceedings reflects the collective interests and judgments of the scholars
who responded to the call for abstracts for the Conference. Roughly half of the papers focused on
factors that influence attitudes toward integrity and actual research practices. These factors are
explored in these papers from the perspective of students and mentors, institutions and professions,
medical practice and clinical research, conflict of interest, and, the most-studied subcategory of
integrity, research misconduct. A second group of papers looked specifically at the way research
integrity is taught, either across institutions or in one institution or course. Finally, a significant
number of scholars tackled important methodological issues, looking at specific ways to detect
misconduct, publication practices, and different theoretical perspectives.
To speed dissemination and to facilitate access, all of the papers published in this volume have
previously been made available on the web. This limited-edition, bound copy is intended to create a
more permanent archive of the first Research Conference on Research Integrity. As this volume goes
to press, the call for abstracts for the second Research Conference on Research Integrity is being
transmitted to continue the work begun in November 2000.

Nicholas H. Steneck, Ph.D.

Department of History, University of Michigan
Office of Research Integrity, DHHS

Mary D. Scheetz, Ph.D.

Office of Research Integrity, DHHS

vii
viii
 Contents
Preface v
Introduction vii
Assessing the Integrity of Publicly Funded Research 1

I. Norms and Environmental Issues

1. Students and Mentors 17

What Would Get You in Trouble: Doctoral Students’ Conceptions of Science and Its
Norms (19)
Data Manipulation in the Undergraduate Laboratory: What are we teaching? (27)
Preliminary Observations on Faculty and Graduate Student Perceptions of Questionable
Research Conduct (35)
Constructing a Personal Model of Research: Academic Culture and the Development of
Professional Identity in the Professorate (41)
Undergraduate Academic Cheating as a Risk Factor for Future Professional
Misconduct (49)
2. Institutions and Professions 55
Comprehensive Guidelines for the Responsible Conduct of Researchers (57)
Research Integrity in Social Work: Status, Issues, and Future Directions (65)
Organizational Influences on Scientific Integrity (73)
3. Medical Practice and Clinical Research 91
Waiving Informed Consent: Long-Term Consequences for the U.S. Military (93)
Promoting Scientific Integrity: The Long Road Ahead–Some Considerations from
Espírito Santo, Brazil (99)
Ethical Research Practice with Human Participants: Problems, Procedures, and Beliefs of
Funded Researchers (105)
Balancing Risks and Benefits of Deception in Assessing Genetic Screening (113)
Research Integrity and the Direct Involvement of Persons with Disabilities (117)

ix
4. Conflict of Interest 125
What is Driving Policies on Faculty Conflict of Interest? Considerations for Policy
Development (127)
The Commercialization of Academic Science: Conflict of Interest Policies and the
Faculty Consultant (133)
5. Understanding Misconduct 141
Preventing Scientific Misconduct: Insights from “Convicted Offenders” (143)
The Relative Efficiency of Research Misconduct Investigations Involving Personal Injury
vs. Injury to the Scientific Record (151)
Ethical Evaluation of Misconduct Cases (159)
Potential Cultural Factors In Scientific Misconduct Allegations (163)
Whistleblowers in Environmental Science, Prevention of Suppression Bias, and the Need
for a Code of Protection (167)

II. Teaching

6. Training in the Responsible Conduct of Research 177

Influencing the Moral Dimensions of Professional Practice:
Implications for Teaching and Assessing for Research Integrity (179)
Research Ethics in US Medical Education: An Analysis of Ethics Course Syllabi (189)
Teaching Ethics in Biomedical Science: Effects on Moral Reasoning Skills (195)
Fostering Research Integrity through Educational Programs: Lessons Learned at the
University of Minnesota (203)
Being a Scientist: Educating for Ethical Conduct (209)
Encouraging Accountability in Research: A Pilot Assessment of Training Efforts (215)
A Plea for Pursuing New Dimensions of Assessment in the Teaching and Learning of
Research Integrity (223)
7. Responsible Conduct of Research Courses 229
The Responsible Conduct of Animal Research (231)
An Effective Short Course on Research Integrity (237)
Resources for Instruction in Responsible Conduct of Research (241)
An Interactive Web Site for Ethics Training: http://storiesandquestions.com (247)

III. Research Theory and Methods

8. Detection Methods 253

The Misuse of Statistics: Concepts, Tools, and a Research Agenda (255)
Images as Evidence: Forensic Examination of Scientific Images (261)
Terminal Digits and the Examination of Questioned Data (269)

x
9. Publication Practices 279
Guidelines on Plagiarism and Paraphrasing in Writing Manuals Across Various Disci-
plines (281)
Instructions to the Author: An Integrity Issue (285)
Erratum Citation and Accuracy in the Publication Record (291)
10. Theory and Models from other Disciplines 297
An Epistemic Model for Moral Hazards in Scientific Enterprises (299)
Scientific Misconduct as Organizational Deviance (305)
A Market Approach to Research Integrity (315)
Methods for Research on Research Integrity: Doing Research on Sensitive Topics (321)
Research Misconduct: A Multiperspectival Approach (327)

Keyword Index 335

xi
Assessing the Integrity of Publicly Funded Research
Nicholas H. Steneck, Department of History, University of Michigan, USA

Keywords Accuracy, Authorship, Bias, Conflict of interest, Duplicate publication, Misconduct, Other
research practices, Peer review, Research on research integrity, Self correction

Since the early 1980s, when research integrity became a major national concern as a consequence of
reports of misconduct in research, several thousand publications have in one way or another reported
on, analyzed, and/or expressed opinions about the integrity of publicly funded research. Despite
widespread interest in research integrity, however, the integrity of researchers has not been subject to
the same critical study as other professionals. The research articles listed at the end of this paper
account for no more than 3-4% of the total literature on research integrity.
The lack of research on research integrity presents a significant problem for government, research
institutions, and professional societies. If integrity is defined as being honest in your dealings with
others, there is ample evidence to suggest that from time to time publicly funded research falls short
of this mark. As the articles summarized in this Paper confirm, researchers do commit misconduct;
research results are inappropriately influenced by bias, conflicts of interest, and just plain
carelessness; and researchers allow personal ambitions and biases to get in the way of the supposed
objectivity of the research process. Publicly funded research does not always achieve the high
standards that researchers, research institutions, and professional societies commonly set for
themselves. This much is known.
In contrast, too little is known about the causes and significance of, or remedies for, research
practices that fall short of the ideals set for the responsible practice of research.
• Is research misconduct rare or are the cases reported simply the tip of some unmeasured iceberg?
• Are there accepted norms or standards for research and, if so, how are they set, learned, and
monitored?
• Are the regulations that currently govern publicly supported research sufficient and well enough
enforced?
• Which practices that seem to fall short of accepted standards matter most from the standpoint of
protecting the public’s investment in research?
• Are there ways to foster integrity and thereby to prevent misconduct?
• Do research ethics courses make any difference?
• What influence does the research climate have on research integrity?
Each of these questions has at one time or another been raised and answered in the literature on
research integrity. Few of the answers given have been based on critical understandings of research
The information and views presented in this report are those of the author and do not reflect the official views or policies of
the Office of Research Integrity or the co-sponsoring organizations.
Corresponding author: Nicholas H. Steneck, Department of History, University of Michigan, Ann Arbor, MI 48109-1003,
734-647-4868 (voice), 734-647-4881 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
as a profession, largely, as noted, because a measure of the degree to which researchers
research as a profession has not be the subject of adhere to the rules or laws, regulations,
careful observation and controlled study. guidelines, and commonly accepted professional
The remainder of this Paper presents a brief codes and norms of their respective research
analysis and summary of the research literature areas.
on research integrity. Finally, a note of caution needs to be added.
• Section one presents an overview of what is This survey of the RRI literature is of necessity
known about the frequency of research selective and evolving. It places more emphasis
misconduct (FFP). on the biomedical sciences than the physical or
social sciences. It does not do justice to the rich
• Section two discusses the complex and
literature on peer review. It almost certainly has
growing literature on research practices that missed important articles that need to be included
seemingly compromise professional stan- in the RRI literature. As a result, it will almost
dards but may not constitute outright mis- certainly be updated, and therefore comments
conduct. and additions are welcomed.
• Section three surveys the research that has
been done on approaches to providing Misconduct
instruction on the responsible conduct of Opinion about the extent of misconduct (FFP) in
research (RCR). publicly funded research is sharply divided. In
• Section four explains how the literature cited public testimony and editorials, researchers have
in this Paper was selected, some of its commonly argued that research misconduct is
characteristics, and the limitations of this rare. Support for this position is based on the
fact that the documented cases of misconduct are
analysis.
few in number in comparison with the total
The bibliography at the end provides a complete
number of individuals engaged in research.
list of references cited in the Paper, a summary of
Approximately 200 cases of misconduct have
the RRI literature sorted by topics, and a
been confirmed by the federal government over
comprehensive listing, sorted by first author, of
the last decade. Dividing cases by total
the RRI literature with abstracts. researchers, this works out to a rate of about 1 in
Throughout this Paper, I have used the terms 10,000 over 20 years, assuming approximately
“research misconduct,” “scientific misconduct,” 2,000,000 active researchers, or 1 in 100,000 per
or simply “misconduct” to refer to the three year. Critics of the way publicly funded research
behaviors outlined in the common government is conducted and administered counter that the
definition of research misconduct, namely reported cases represent the tip of a larger but
fabrication, falsification,and plagiarism (FFP) in uncharted iceberg. Support for this view is based
proposing, conducting or reporting the results of in part on documented and presumed examples of
research. While none of these behaviors is self- the reluctance of researchers and research
explanatory, the crucial element in each is a institutions to pursue cases of misconduct (for
deliberate intent to deceive or mislead. Delib- early warnings about possible larger numbers,
erate deception is clearly not consistent with see: 1, 2). Which, if either, opinion is correct
good research practice and is generally agreed to remains to be determined.
constitute misconduct.
A second term used throughout this report, Direct evidence
“integrity,” is more difficult to define. Integrity
Research undertaken to clarify the extent of
is a measure of wholeness or completeness.
scientific misconduct suggests that it may be
When applied to professional behavior, it is
more common than the 1 in 10,000 or lower
essentially a measure of the degree to which
estimates. Evidence for this position comes from
someone’s (or some institution’s) actions accord
three direct approaches to measurement:
with ideal or expected behavior. However, the
ideals or expected behaviors for professional • It is reasonable to presume, based on research
conduct are complex, not always well defined, in other fields, that confirmed cases underes-
and subject to change or reinterpretation. I have, timate actual cases (3). Further research is
therefore, adopted a fairly inclusive definition of needed to determine whether under-reporting
integrity and assumed that it can be thought of as in research is trivial or significant.
2
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research

Year Population Sample Responses Mis- FFP

Author Place Size (%) conduct

1976 Readers, New Scientist ?? 199 92% ?

St. James-Roberts En gland (?)
1987 Phys, biol, behav, & soc. scientists 1100 245 – 32%
Tagney Major research university, US (22%)
1992 Biomedical trainees 2010 549 36% –
Kalichman UC San Diego, US (27%)
1993 Chem., civil eng., microbiol., sociol. 4000 --/-- 44/50% 6/9%
Swazey US survey, faculty/graduate (72/59%)
1993 PIs, biomedical sciences 159 119 27% –
Hals Health Region IV, Norway (70%)
1995 Biomedical researchers 274 215 22% 3%
Bekkelund Norway, random survey (80%)
1996 Post-doctoral training fellows 1005 324 58% 3-12%
Eastwood US, random national survey (33%)

Table 1. Surveys of the Level of Misconduct in Research

• Surveys of knowledge of misconduct consis- troubling discrepancy between public statements

tently report knowledge rates above 1% about how “rare” misconduct in research
(Table 1). Reported knowledge of miscon- supposedly is and the more private belief on the
duct remains above 1% (1 in 100, or 100 part of many researchers that it is in fact fairly
common. How can these two views be
times higher than the 1 in 10,000 estimate)
reconciled?
even when researchers are asked about their
Second, whatever the actual rate of
own research group and when misconduct is misconduct, it is not so much the rate as the
specifically limited to FFP. One survey significance of the misconduct that matters most.
specifically asked researchers whether the Summarizing the results of scientific data audits
misconduct they were aware of was public of the Cancer and Leukemia Group B’s clinical
knowledge. Of the roughly one-in-four trials, Weiss et al. conclude that “scientific
researchers who were aware of misconduct improprieties have occurred very rarely...”
(27%), 47% said that the cases were not (8, p. 459). “Very rarely, in this case, is based on
public knowledge (4). a quantitative estimate of 0.28% (p. 462)–28
• Audits of research procedures and results have cases of misconduct for every 10,000 clinical
researchers or one case for every 357 clinical
turned up “significant problems” or “major
researchers. On what basis can this rate be
deviations” a levels that range at and above judged as either “rare” or “significant”? Clearly,
the 10% level (5-8). These results do not understanding the importance of misconduct in
correlate directly with FFP, since they do not research requires not only better estimates of
take into account whether discrepancies numbers but also of significance. How much
result from deliberate actions. does a case of misconduct in research actually
The results of surveys, audits, and estimates of cost the public in terms of wasted research
the rate of under-reporting raise two important dollars, of deceptive findings that mislead other
issues for further consideration. First, however researchers until the misconduct is discovered,
the results of surveys and audits are ultimately and perhaps of negative impacts on patient
interpreted or clarified, there remains the health?

3
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Indirect evidence career pressures may make it necessary to engage

Gathering information on the likely prevalence of in practices that they also know are wrong.
misconduct in research can be approached That significant numbers of beginning
indirectly. For example, many studies have researchers may in fact do what they say they
documented that cheating is common in the will do has been confirmed in a series of audits
educational system at all levels and in all of the research publications listed on residency
programs. The rates vary from well above 50% fellowship applications. These audits report
for high school and college undergraduates (9- significant numbers (15% and higher) of
12) to levels between 10% and 30% for misrepresentations, from seemingly trivial
professional students (13-20). One survey offenses such as inflating author rank to listing
specifically asked whether misconduct at this articles “in press” when they were not, listing
level was indicative of future performance. Of papers in journals that do not exist, and listing
246 faculty and administrators responding, 216 bogus articles in real publications (Table 3) (23-
(86%) felt that it was so indicative (14, p. 34). If 27). Similar practices are generally counted as
this estimate of the relationship between student FFP when they occur in research grant
conduct and later professional conduct is true, it applications or resumes submitted for promotion.
would support the contention that the prevalence One final piece of indirect evidence that
of misconduct in research may be higher than the should be noted is the confirmed reluctance of
small number of confirmed cases suggest. researchers to report suspected misconduct.
The prevalence of a willingness to engage in • As noted above, Hals reported that roughly
misconduct has been documented into graduate one-in-four researchers (27%) who knew of
and post-doctoral research education. misconduct, said that the cases they knew of
Kalichman’s and Eastwood’s surveys report that were not public knowledge, which could
significant numbers of students (above 10%, mean they were not reported (4).
except for fabricating data) will omit or change
• In Tagney’s survey conducted at one research
evidence and add honorary authors if it will help
get papers published or grants funded (Table 2) institution, roughly half of those who re-
(21, 22). Students who are in the beginning ported suspecting misconduct took no action
stages of becoming researchers clearly feel that (28).
• Korenman’s
1992 1996 study of the
Action
Kalichman Eastwood attitudes of
Past misconduct (yes/no?) 15.1% 12% researchers and
Future misconduct (yes/no?) 14.8% institutional
...modify data for paper 7.3% 15% representatives
...modify data for a grant application 13.5% -- toward miscon-
...fabricate date for a paper or grant application 1.3% < 2% duct found that
...select or omit data for paper or grant application 14.2% 27% researchers were
...list an undeserving author -- 41% more likely to
favor informing
Table 2. Self-reported attitudes toward misconduct

Author 1995 1996 1997 1998 1999

Sekas Gurudevan Panicek Bilge Dale
Specialty Gastro- Emergency Radiology Pediatrics Orthopaedic
enterology Medicine Medicine
Total applications 236 350 201 404 213
...with citations 53 (22%) 113 (32%) 87 (43%) 147 (36%) 64 (30%)
...misrepresented 16 (30%) 23 (20%) 14 (16%) 29 (20%) 11 (17%)
Total citations -- 276 261 410 76
...misrepresented -- 44 (16%) 39 (15%) 41 (10%) 14 (18%)
Research experience 138 (59%) -- -- -- --
...not confirmed 47 (34%) -- -- -- --
Table 3. Misrepresentation in medical resident training program applications

4
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research

colleagues whereas institutional representa- Panel on Scientific Responsibility and the

tives favored reporting to supervisors and Conduct of Research specifically set out a
deans (29). separate category of research behavior called
These findings confirm the suspicions of the “tip- “Questionable Research Practices.” The Panel
of-the iceberg” school, which argues that recognized that such practices “...violate
reported cases are not an accurate measure of traditional values of the research enterprise and
actual levels of misconduct. No controlled ... may be detrimental to the research process,”
studies of under-reporting have been undertaken but it was not willing to include them under
to assess the rate of under-reporting, making it “misconduct.” It did concede, however, that
difficult to conclude whether it is significant. since “...the relationship between these two
Cheating or misconduct on the path toward categories is not well understood ... [i]t may be
becoming a researcher does not, of course, difficult to tell, initially, whether alleged
demonstrate that misconduct continues once misconduct constitutes misconduct in science or
students become researchers. Under-reporting a questionable research practice” (30, pp. 5-6,
may not seriously compromise estimates of the 29).
amount of misconduct. Reasons can be given to Whether or not “other questionable
suggest that some of the estimates of misconduct practices” constitute misconduct is irrelevant for
given in the various surveys reported above may the purposes of this Report. What is relevant is
be too high as well as reasons to suggest that they the fact that any practice that deviates
may be too low. The differences between the significantly from the “rules, regulations,
“rare” and “tip-of-the-iceberg” schools can guidelines, and commonly accepted professional
therefore not be resolved easily. What is codes or norms for the responsible conduct of
important to note, however, is that in seeking to research” (the definition for integrity given in the
refine understandings and resolve the differences Introduction) can compromise and currently are
between the two schools, the range of uncertainty compromising the integrity of publicly funded
that exists is significant. In terms of decimal research. However, until more is known about
points, the range is not a matter of one or two these practices, it will be difficult to suggest how
orders of magnitude but closer to four or five seriously they need to be taken.
orders of magnitude, varying from 1 in 100,000 The remainder of this section summarizes
or less to 1 in 100 or more. And this, in turn, some of the research on other practices that can
makes it difficult, if not impossible, to estimate compromise the integrity of research. The
the public costs of misconduct when determining summary is intended to be more illustrative than
what policies are needed to protect the public’s exhaustive. Some aspects of research practice,
investment in research. such as authorship and peer review, have been
the subject of intense study and hundreds of
Other Research Practices publications, thanks in large part to the
Over the past twenty years or longer, the Congresses on Biomedical Peer Review
discussion of “research integrity” has focused organized by JAMA editor, Drummond Rennie
primarily on “research misconduct,” based on (31). Exhaustive coverage is therefore not
widespread agreement that misconduct (FFP) is possible. Rather, the goal of this section is to
wrong or fraudulent. While it is true that focus on some areas of potential concern and
research misconduct clearly can undermine the illustrate some of the findings that have emerged.
integrity of publicly supported research and
therefore needs to be taken seriously, so can other Accuracy
research practices, such as sloppy research, Accurate information is vital to research.
inappropriate bias, conflict of interest, or poor Research is a cooperative and cumulative
mentoring. enterprise. Researchers build on the work of
The existence of other research practices that others, which means the information they have
can compromise integrity has been recognized by about other work and the way research is
the research community, but there has been no conveyed must be accurate; however, a number
agreement on how to respond to them or how of studies suggest that research results are not
seriously they should be taken. In its 1992 always conveyed accurately.
report, Responsible Science, the NAS/NAE/IOM • Information presented in abstracts does not

5
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

always accurately reflect the information substantial agreement about whether to

given in the article itself. One study reported publish (36) and by studies of how fraudu-
major discrepancies in abstracts (inconsisten- lent publications have made it to press (37).
cies or information that was not contained in How much effort should be made to improve
the body of the article) in 55 of 203 ran- peer review requires more information about
domly selected articles (32). how well it is working and the price of its
• Studies have reported that significant numbers shortcomings.
(above 10%) of published articles misuse • Peer review has been shown to have institu-
statistics or contain statistical errors (33). tional (38), national (39, 40), methodological
• Random checks on citations and quotations in (39, 41), gender (42) and outcome biases
published articles have reported error rates (43-45). Bias, obviously, runs counter to the
well above 10%. Errors were counted as value-neutral goal of research.
“citation errors” when the names, pages, or • Considerable uncertainty exists about the best
other information needed for locating an ways to improve peer review. Traditional
article was inaccurate (minor) or when the approaches, such as blinding, issuing clear
referenced article could not be located based instructions, or relying on experienced
on the information given (major). Errors researchers, have had different measures of
were counted as “quotation errors” when the success (46-53).
reference oversimplified or exaggerated • Studies of peer review have raised questions
information given in the referenced article about whether it helps or hinders innovation
(minor) or when the information given in the (54, 55).
original article did not support or contra- One review of the rich literature on peer review
dicted claims made in the reference (major) concludes: “Because of the central place of peer
(34, 35). review in the scientific community and the
Inaccuracies in abstracts, the use of statistics, and resources it requires, more studies are needed to
references do not necessarily invalidate research define what it does and does not accomplish”
results. Conclusions or pieces of evidence (56). This work will fortunately be fostered by
presented only in an abstract but not in the body the future Congresses on Biomedical Peer
of an article could be true. Research results Review and similar efforts.
bolstered by inflated or deceptive statistics or
inaccurate references to other studies might still Self-Correction
be true. At issue, however, is not whether the Researchers constantly read and check each
results are ultimately true or accurate but whether other’s work. The routine process of using the
the word (or words in this case) of researchers work of others in the day-to-day practice of
can always be trusted. The clear answer to this research provides an additional mechanism for
question, unfortunately, is that it (they) cannot. detecting and correcting errors and other
problems in research, such as research
Peer Review misconduct. Research is, in other words, self-
Inaccuracy and other problems in publication are correcting, which further ensures its integrity.
purportedly reduced, if not eliminated, through However, research on the effectiveness of self-
peer review. In general, the peer review system correction in research has shown that this
enjoys considerable support within the research mechanism is not as vigilant as one might expect.
community and is seen by most as the foundation • Studies of some of the first publicly docu-
on which professional self-regulation rests. This mented cases of misconduct found that
does not mean, however, that peer review is publication of a retraction reduced the
above criticism or not in need of further citation of fraudulent articles but did not
improvement. eliminate it (57-59).
• That peer reviewers miss problems in publica- • One recent study of articles retracted for a
tions has been documented by the fact that broad range of reasons, from outright fraud
different reviewers detect different problems to acknowledged experimental errors or later
in manuscripts, even when they are in failure to replicate, concluded that retracted
6
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research

articles continue to be cited and used as a for judging authorship found that 19% (65)
significant rate. Of 299 post-retraction and 36.4% (66) of papers did not meet these
citations listed in the Abridged Index criteria.
Medicus, only 19 (6%) mentioned the • Evidence suggests that the rules for author-
retraction; 17 (6%) explicitly and 263 (88%) ship are poorly understood, interpreted
implicitly reported the retracted work as differently by different researchers, and not
“valid” (60). well communicated from senior to junior
• Research on the process by which articles are researchers (22, 67, 68).
retracted and erroneous information with- • Patterns of authorship and the increase in
drawn has show that it is slow (60, 61) and disputes over authorship suggest that deci-
in some key ways ineffective (60-63). sions about authorship are significantly
Findings such as these have important policy influenced by the research environment (69,
implications. In his study of retraction notices, 70).
Budd agrees that research is self-correcting, but The importance of the truthful reporting of
then he adds: “...there may be a great deal of research contributions through authorship is
time, effort, and money spent in discovering that widely recognized. The NIH Guidelines for the
some research is not useful. If erroneous or Conduct of Research note in particular that:
fraudulent work lives on in the literature, the For each individual the privilege of authorship
amount of time, effort, and money to correct should be based on significant contribution to
work may be even greater” (60, p. 297). At the conceptualization, design, execution, and/
issue, in other words, is not whether research or interpretations of the research study, as well
errors are corrected, but when. Failure to correct as a willingness to assume responsibility for
the literature in a timely and responsible manner the study. Individuals who do not meet these
is as much a matter of integrity, viewed from the criteria but who have assisted the research by
their encouragement and advice or by providing
public’s investment in research, as a failure to
space, financial support, reagents, occasional
correct at all. analyses or patient material should be
acknowledged in the text but not be authors.
Authorship (71, p. 10)
In principle, research results are more important Authors who ask or agree to be listed on papers
than researchers. Who publishes an article to which they have not made substantial
should not matter. What matters most are the contribution compromise the integrity of the
results. In practice, however, authorship is research environment. The same is true of the
vitally important to, and significantly influences, 41% of graduate students who report a
the research process. Most research funding willingness to list undeserving authors on their
today is dependent on productivity. Review papers (see Table 3, above).
panels want to know not only what a researcher
is planning to do but what she or he has done. Duplicate Publication
Advancement in academic research is not In its advice to intramural researchers, NIH
possible without publication. Getting one’s name research Guidelines caution researchers about
on research papers is important–so important that duplicate publication:
as many as one in five aspiring researchers Timely publication of new and significant
misrepresents publications on résumés in an results is important for the progress of science,
attempt to improve his or her standings as a but fragmentary publication of the results of a
researcher (see Table 4). scientific investigation or multiple publications
As with the other research practices of the same or similar data are inappropriate.
discussed in this section, there is considerable (71, p. 8)
evidence to suggest that the ideal standard for Despite widespread agreement that duplicate
determining authorship is not followed in publication is inappropriate, the rate of duplicate
practice and that expected authorship practices in publication (publishing the same article twice
general are sometimes not clearly defined or without reference) seems to hover at about 10%
conveyed. (Table 4) (72-76). Based on his study of
publication trends in the British Medical Journal,
• Two studies that used the ICMJE criteria (64)
Waldron suggested that duplicate publication was
7
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Study Journal Articles Duplicate %

Waldron (1992) BMJ 354 published 6-12%
Bernard (1993) NTvG 172 published 11%
Koen (1994) NTvG 108 rejected 4%
Blancett (1995) INJS 642 published 9%
Bloemenkamp (1999) NTvG 148 published 7%

Table 4. Percent duplicate publication

increasing (72). Bleomenkamp more recently tenets” (p. 138). From the public’s perspective,
reported that the duplicate publication rate for however, it makes no difference whether the
articles in Nederlands Tijdschrift voor duplication is intended or not. If researchers do
Geneeskunde has remained constant over the last not take steps to ensure that a second or third
10 years and the number of authors referencing publication of a body of data is recognized as
the second publication has increased such, the public could be harmed and the
significantly, from 22% to 73%.(76). integrity of the research process undermined.
Duplicate publication adversely effects
research in a number of ways. It can waste time Bias and Conflict of Interest
(editors and reviewers) and resources (library There has been considerable debate about the
funds and reprint costs). It also makes it difficult role of values and personal interest in research
to evaluate the productivity of researchers. But ever since Merton proposed “disinterestedness”
perhaps most importantly, in clinical research it as one of four key values on which science rests
has the potential to inappropriately distort or (79, p. 116). It is now widely recognized that
bias findings if the duplicate publications are values influence research (80), but there is also a
more prevalent in one treatment regimen. common understanding that the influence of
• In a meta-analysis of post-operative effects of values should be minimized and made public,
ondansetron, Tramer and Reynolds reported particularly when financial interests are involved.
that “17% of published studies and 28% of Considerable evidence exists to support the
the patient data were duplicated. Moreover, contention that personal interest does influence
research behavior. Positive-outcomes bias
duplication was more common in studies that
(favoring publications that report positive results
reported greater treatment effect. This bias, over those that report negative results or that do
according to Tramer and Reynolds, “led to a not find results) has been demonstrated in a
23% overestimation of ondansetron’s number of studies (44, 81, 82). The reverse
antiemetic efficacy” (77). effect has also been reported, that is, slower
• Jefferson reports that in a Cochrane review of publication rates for studies that fail to find a
the effects of Plasma Derived Vaccines, he particular result (45). Studies are just beginning
and his colleagues suspected that 25% (15 of to assess how these interests affect research and
60) of the trials identified during the first whether they are being properly managed (83-
phase of review were duplicate publications. 85).
In calling controversial publication,
This percentage increased to 43% (3 of 7)
reporting, and other research practices
when they progressed to the second phase of “questionable,” the NAS report, Responsible
review. Being aware of the problem of Science, highlights an important problem. (30)
duplicate publication, his group excluded the “Integrity” is not an all-or-nothing proposition.
duplicate studies, but doing so is not com- There is a difference between a failure to check
mon practice (78). the spelling of every author’s name or to catch
In the final analysis, Jefferson considers only every typo and using improper statistics or
“publishing redundant material with the intention delaying the publication of a manuscript to please
of misleading the public, editors and readers, in a sponsor. It is not easy to pinpoint where or
order to make them believe the study is different when high standards for integrity in research give
from the original” as a “breach of current ethical way to careless research practices, to
8
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
irresponsible research practices or to misconduct. succeeding.
The extremes (high standards for integrity and
misconduct) can be defined, but behaviors that RCR training
fall between, to one extent or another, are all Straightforward efforts to evaluate the impact
subject to interpretation. This, in turn, makes it RCR training has on attitudes or anticipated
imperative that these behaviors are well behaviors have not reported any clear positive
understood and their consequences evaluated, results. Studies by Kalichman et al. and
both as part of the process of reassuring the Eastwood et al. compared receiving or not
public that its research funds are being spent receiving RCR training with anticipated research
responsibility and as needed background behaviors. A study by Brown compared
information for developing responsible conduct receiving or not receiving RCR training with
of research training programs. self-reported perceptions of different ethical
standards. None of the studies found any
Education significant correlations between attitudes or
It is commonplace for reports on research anticipated behaviors and RCR training (21, 22,
misconduct/integrity to emphasize the 88). Brown’s study did report that RCR training
importance of education. Professions have an increased awareness of options in ambiguous
obligation to society to educate future situations (p. 490). However, Eastwood’s study
generations of professionals, which includes reported that fellows who received RCR training
making future professionals aware of the were more willing to grant honorary authorship
standards for responsible practice. Moreover, if than fellows who did not (p. 95). Overall, direct
professional ethics education prevents measures of attitudes and anticipated behavior
misconduct, it is in a profession’s best interest to have pointed to some possible benefits, perhaps
encourage this education, which most in fact do. one puzzling negative, and a great deal of
Through the 1980s, research ethics training similarity between those receiving and not
was commonly relegated to the laboratory and to receiving RCR training.
mentoring. This changed in 1989 when NIH and Efforts to refine the study of the impact of
ADAMHA instituted required “instruction in the RCR training have led to a difference of views on
responsible conduct of research” (RCR) for all appropriate outcome measures. Based on a
training grants (86). The requirement stipulated three-year effort to develop and assess an RCR
that training programs had to have instruction in course at Dartmouth College, Elliot and Stern
RCR, which in turn had to be described in the argue that “if ‘ethical behavior’ is removed as a
training grant application. Although the basis for the evaluation of teaching ethics,”
requirement technically had no “regulatory effective assessment tools can be developed. In
teeth,” coming as it did in the highly competitive the place of ethical behavior, they propose using
environment of grant-getting, researchers and two familiar measures of success in academic
research institutions quickly complied and courses in general: “the skills and content taught
instituted a wide variety of research ethics or in the course and the learning environment in
RCR training programs (87). which the teaching takes place” (89, p. 348). The
The increase in formal RCR training raises project allowed them to develop and test various
an obvious and researchable question: has it or tools for evaluating these ends, which they argue
will it make any difference? At the present time, can be accomplished, “but only if [teaching of
there is no convincing evidence that it does, but academic research ethics] is treated as an
this does not necessarily lead to the conclusion academic discipline by both faculty and students”
that RCR training is ineffective, unnecessary, or (p. 355).
unwise. The newness of most programs means Others believe that striving for some type of
that their impact may not yet be apparent. RCR behavioral or moral reasoning change is
training is delivered in different ways and appropriate for professional ethics instruction,
different settings, making it difficult to isolate the including RCR training, and that such change can
influence this one factor has on the complex be measured. In a series of studies of medical,
process of becoming a responsible researcher. veterinary, and dental education, Self, Baldwin,
And perhaps most importantly, there is no Bebeau and colleagues have reported that: a)
agreement on the goals of RCR education, traditional professional education programs may
making it difficult to judge whether it is erode and b) the addition of ethics instruction to
9
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
traditional programs improves the ability of there was room for improvement (99, 100).
students to engage in moral reasoning (90-97). Junker suggest that more journals should
Whether changes in the ability to engage in require authors to follow the Consolidated
moral reasoning measured in professional Standards of Reporting Trials (CONSORT)
education settings generally can be applied to
(101). Clarke and Chalmers conclud that
RCR training in particular and whether changes
in moral reason have any lasting professional “there is little evidence that journals have
consequences remains to be determined. adequately implemented the CONSORT
The research needed to plan effective RCR recommendation that results of an RCT
programs will clearly need to take into account [randomized controlled trial] be discussed in
more than what goes on in the RCR classroom. light of the totality of the available evidence”
Studies have shown that environment is closely (p. 280).
linked to what students feel they must do as • In studies of measures to improve the quality
opposed to what they should do (17, 18, 20, 22). of abstracts, Pitkin found that instructions to
Although the 1995 survey of the attitudes and the authors had little impact (32, 102, 103).
experiences of 2,000 graduate students with
• In a study of the impact of guidelines pub-
misconduct (Table 2, above) indicates “that
fraud, plagiarism, and related forms of lished in the British Medical Journal for
misconduct are the results of individual manuscripts on the economics of health care,
predilections or failures of judgement...” (98, p. no difference was found in the quality of
225), Anderson et al. in commenting on these manuscripts, although the guidelines were
results still point to important influences exerted judged to be useful for editorial purposes
by environment and mentoring relations (p. 226). (104).
Without attention to the full context within which • In a comparison of systematic reviews and
integrity is learned and decisions made about meta-analyses published following the
right and wrong actions, the goal of ensuring the procedures of the Cochrane Collaboration
responsible conduct of research through RCR
versus the more open-ended general reviews
training could well be negated by influences in
the research environment. published in journals, Jadad reported more
methodological rigor in the Cochrane
Other efforts to educate reviews (41).
In discussions of ways to improve the integrity of • In a study of the impact of professional codes
research, surprisingly little attention has been in physics, Tarnow reported that postdoctoral
given to the role of clear rules and routine students were generally not aware of publi-
monitoring or data audits. If the ultimate goal of cation rules and spent little time with advi-
research ethics/integrity policy is simply to sors discussing publication practices (68).
ensure high standards for publicly supported As a group, this research seems to support the
research, the simplest way to achieve this goal perhaps not unexpected conclusion that rules
may be to make the rules as explicit and clear as alone will not change behavior and must be
possible and then to check to make sure they are accompanied by efforts to both make them
being followed. For each of these approaches to known and take them seriously. Simply making
“educating” researchers, there is interesting information about rules for responsible behavior
research that suggests what may or may not available is not an effective way to foster
work. responsible behavior.
Over the last decade, new rules have been In contrast, data audits seem to have a
formulated for reporting research. Particular significant effect on research behavior. Two
attention has been paid to two key areas–journal studies of major government data audit programs
publication in general and clinical trial reporting. both report that serious misconduct declined over
Studies of the effect of new rules suggested that the course of the studies.
they have had mixed results. • Shapiro and Charrow’s study of FDA audits
• Two studies that looked at the adoption of conducted between 1977 and 1988 reported
specific standards for reporting clinical trials that the rates of specific deficiencies re-
by several medical journals concluded that mained about the same throughout but “the

10
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research

overall level of seriousness of the problems research integrity is based on some research.
... declined” (7, p. 130). For the purposes of this Report, “research”
• Weiss et al. in their detailed look at the results has been defined as studies that have some
of audits conducted by the Cancer and element of controlled investigation, which means
primarily but not exclusively surveys and
Leukemia Group B (CALGB) conclude that:
quantitative assessments. Limiting the definition
“The CALGB data audit process has been of research in this way obviously eliminates
successful in uncovering the very rare many thoughtful articles and books from the
instances of scientific misconduct and literature review, such as editorials, analytical
pressuring group members to improve writings, historical and cases studies, and
adherence to administrative requirements, philosophical analyses. The fact that works such
protocol compliance, and data submission. It as these are not included in this Report should
has also served to weed out poorly perform- not be taken as suggesting they are not important.
ing institutions” (8, p. 464). They clearly are crucial and in other contexts
If results matter, then one of the most effective certainly need to be considered. However, for
ways to educate researchers about their the purposes of the ORI RRI program, the
responsibilities may be to check more carefully immediate goal is to gather hard evidence
the work they produce. relating to actual research practices, so that
Data audits have been resisted because they policy-making can be based on the way research
are allegedly expensive, time-consuming, and is conducted as opposed to the way we may think
perhaps even counter-productive; e.g. too much it is conducted.
concern about the bookkeeping required to pass Controlled quantitative research plays an
audits might slow the progress of science. There important role in scholarly investigation. Most
currently are no data to support these concerns. significantly, it helps establish reference points
There is evidence, reviewed by Armstrong, that for organizing and evaluating other information.
peer review can slow innovation in research (54, For example, historians, journalists, and others
pp. 70-71), but no evidence that data audits have have amply documented that misconduct takes
a similar effects. Moreover, Glick’s rough place in research. However, without some
estimates of the cost of data audits, based on quantitative assessments, it is difficult to know
conservative estimates of the amount of careless what to make of individual cases of misconduct
work and misconduct that may be affecting or even of the entire body of confirmed cases.
research results, suggests that over the long term, Are they typical or atypical? Is misconduct
they will save public dollars. “Data auditing common or rare? Without some controlled
would increase research productivity by 2.5-6% counting or surveys, it is difficult to place
(...), so that each dollar spent on such audits individual events and behaviors into context.
might eventually benefit the public, 20 years Locating research on research integrity is not
later, by an amount equivalent to $25-60” (3, p. a simple task. Keyword searching for the most
81). These results and estimations will no doubt part does not separate scholarly analyses from
be challenged, but for now the evidence seems to empirical studies. References located through
suggest that research audits might be an effective searches for “scientific misconduct,” “research
and efficient way to detect misconduct and ethics” and other keywords need to be evaluated
reduce the rate of other questionable practices. for both relevance and method. The articles
summarized in this Report have been located
Research Literature Overview through standard keyword searches in several
As noted in the Introduction, over the last 20 different databases, checking references listed in
years or longer, several thousand publications bibliographies, and in some cases by searching
have in one way or another addressed the issue of for publications by scholars with known RRI
integrity and/or misconduct in research. Most of interests. Major emphasis has been placed on
these publications are based on some research. work relating to the biomedical sciences in
Reporters do research for news stories. Journal particular and the hard sciences more generally.
editors investigate problems before writing Less attention has been paid to research on
editorials. Taken to mean simply investigation or integrity in the social sciences. The final RRI
study, most if not all that has been written about bibliography contains 136 entries, most of which,
but not all, have some empirical or controlled
11
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
research component. research ethics training and publication
That RRI has not yet developed into an practices. SEE is unfortunately not indexed
organized research field is more than evident in MedLine®, which limits the knowledge of
from the fact that the 136 articles summarized in this important group of publications.
this Report appeared in 45 different journals Together, these three journals account for 76 of
(Table 5) and two books (105, 106). Most the 136 articles. Three journals had three
journals published only one or two articles. research articles; five journals had two, and the
There are, however, three important exceptions. remainder published a single research article on
• Fifty-one of the 136 (37.5%) articles appeared research integrity.
in JAMA. Most of these articles are on The fact that research on research integrity is
integrity in publication and are the product distributed so broadly through the scholarly
of the three peer review conferences orga- literature almost certainly slows research
nized by Drummond Rennie. progress. At the present time, the standard search
• Fourteen of the 136 articles (10%) appeared in tools simply do not cut across the different
Academic Medicine. These articles are disciplines that contribute to RRI. What is
mostly concerned with student conduct, not “discovered” in one field is thus not easily
known in other fields. More importantly,
research integrity specifically, but have been
however, is the fact that the absence of a well
included because they provide important defined literature and corresponding research
background on the values researchers may community makes interdisciplinary research on
have had as students. research integrity more difficult. This second
• Eleven of the 136 articles (8%) appeared in shortcoming is particularly important for the
Science and Engineering Ethics. This group development of research on research integrity,
of publications is split nearly evenly between which of necessity must be interdisciplinary and

Journal of the American Medical Association (51) Cancer Investigation (1)

Academic Medicine (14) Cognitive Therapy and Research (1)
Science and Engineering Ethics (11) Controlled Clinical Trials (1)
British Medical Journal (3) Image: The Journal of Nursing Scholarship (1)
Journal of Professional Nursing (3) Journal of Allied Health (1)
Nederlands Tijdschrift voor Geneeskunde (3) Journal of Bone and Joint Surgery (1)
Accountability in Research (2) Journal of Clinical Epidemiology (1)
Bulletin of the Medical Libraries Association (2) Journal of General Internal Medicine (1)
Journal of Dental Education (2) Journal of Higher Education (1)
Lancet (2) Journal of Information Ethics (1)
Medical Education (2) Journal of Investigative Medicine (1)
Medical Reference Services Quarterly (2) Journal of Medical Education (1)
New Scientist (2) Journal of Medical Ethics (1)
Tidsskrift for den Norske lægeforening (2) Journal of the Am. Veterinary Medical Association (1)
AIDS Education and Prevention (1) Journal of the Royal College of Physicians, London (1)
American Journal of Medicine (1)1 Minerva (1)
American Journal of Public Health (1) Nature (1)
American Journal of Roentgenology (1) New England Journal of Medicine (1)
American Scientist (1) Nordisk Medicin (1)
Annals of Emergency Medicine (1) Nurse Educator (1)
Annals of Internal Medicine (1) Research in Higher Education (1)
Cambridge Quarterly of Healthcare Ethics (1) The Psychological Report (1)
Canadian Medical Association Journal (1)
Table 5. Journals with RRI articles, listed by number of articles.

12
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
broadly inclusive. 8. Weiss RB, Vogelzang NJ, Peterson BA, Panasci LC,
The need for interdisciplinary research raises Carpenter JT, Gavigan M, et al. A successful system of
one last observation about the RRI literature and scientific data audits for clinical trials. A report from the
by implication the RRI community. Most of the Cancer and Leukemia Group B [see comments]. JAMA
1993;270(4):459-64.
literature cited in this Report appears in 9. Stern EB, Havlicek L. Academic misconduct: results of
biomedical journals. The only major exception faculty and undergraduate student surveys. J Allied
are the eleven articles in Science and Engineering Health 1986;15(2):129-42.
Ethics, which, it should be noted, are not indexed 10. Perry AR, Kane KM, Bernesser KJ, Spicker PT. Type A
in MedLine® but are in BioEthicsLine, without behavior, competitive achievement-striving, and
abstracts. That research on the integrity of cheating among college students. The Psychol Rep
biomedical research (the primary focus of this 1990;66(2):459-65.
report) appears in biomedical journals is certainly 11. Schab F. Schooling without learning: thirty years of
understandable, but the existence of this cheating in high school. Adolescence 1991;26(104):839-
47.
publication pattern raises serious questions for 12. McCabe DL. Classroom cheating among natural science
interdisciplinary research. and engineering majors. Sci Eng Ethics 1997;3.
To be taken seriously in most academic 13. Stimmel B, Yens D. Cheating by medical students on
settings today, researchers must first succeed in examinations. Am J Med 1982;73(2):160-4.
their primary research field. This means that 14. Bailey PA. Cheating among nursing students. Nurse
sociologists must publish in sociology journals, Educator 1990;15(3):32-5.
psychologists in psychology journals, and so on. 15. Rozance CP. Cheating in medical schools: implications
In addition, they must pursue research that is for students and patients. JAMA 1991;266(17):2453, 6.
important to their primary fields of research. 16. Anderson RE, Obenshain SS. Cheating by students:
findings, reflections, and remedies. Acad Med
Institutional factors such as this unquestionably 1994;69(5):323-32.
make the development of interdisciplinary 17. Daniel LG, Adams BN, Smith NM. Academic
research on research integrity more difficult. misconduct among nursing students: a multivariate
When added to the fact that there are few investigation. J Prof Nurs 1994;10(5):278-88.
incentives for researchers who are the subject of 18. Baldwin DC, Jr., Daugherty SR, Rowley BD, Schwarz
RRI investigations to study their own integrity, MR. Cheating in medical school: a survey of second-
rather than pursuing research in their primary year students at 31 schools. Acad Med 1996;71(3):267-
fields of interest, establishing an interdisciplinary 73.
RRI initiative and RRI community poses a 19. Dans PE. Self-reported cheating by students at one
medical school. Acad Med 1996;71(1 Suppl):S70-2.
significant challenge. 20. Satterwhite WM, 3rd, Satterwhite RC, Enarson CE.
Medical students’ perceptions of unethical conduct at
Bibliography one medical school. Acad Med 1998;73(5):529-31.
1. Woolf P. Fraud in science: How much, how serious?. 21. Kalichman MW, Friedman PJ. A pilot study of
Hastings Cent Rep 1981;11:9-14. biomedical trainees’ perceptions concerning research
2. Broad WJ, Wade N. Betrayers of the truth: Fraud and ethics. Acad Med 1992;67(11):769-75.
deceit in the halls of science. New York: Simon and 22. Eastwood S, Derish P, Leash E, Ordway S. Ethical
Schuster; 1982. issues in biomedial research: Perceptions and practices
3. Glick JL. On the potential cost effectiveness of scientific of postdoctoral research fellows responding to a survey.
audits. Accountability in Research 1989;1(1):77-83. Sci Eng Ethics 1996;2(1):89-114.
4. Hals A, Jacobsen G. [Dishonesty in medical research. A 23. Sekas G, Hutson WR. Misrepresentation of academic
questionnaire study among project administrators in accomplishments by applicants for gastroenterology
Health Region 4 ]. Tidsskrift for den Norske fellowships [see comments]. Ann Intern Med
Lægeforening 1993;113(25):3149-52. 1995;123(1):38-41.
5. Shapiro MF, Charrow RP. Scientific misconduct in 24. Gurudevan SV, Mower WR. Misrepresentation of
investigational drug trials. N Engl J Med research publications among emergency medicine
1985;312(11):731-6. residency applicants [see comments]. Ann Emerg Med
6. Shapiro MF, Charrow RP. The role of data audits in 1996;27(3):327-30.
detecting scientific misconduct. JAMA 1989;261(May 25. Bilge A, Shugerman RP, Robertson WO.
5):2505-11. Misrepresentation of authorship by applicants to
7. Shapiro MF. Data audits in investigational drug trials pediatrics training programs. Acad Med 1998;73(5):532-
and their implications for detection of misconduct in 3.
science. In: Lock S, Wells F, editors. Fraud and 26. Panicek DM, Schwartz LH, Dershaw DD, Ercolani MC,
miscondcut in medical research. London: BMJ Castellino RA. Misrepresentation of publications by
Publishing Group; 1993. p. 128-41. applicants for radiology fellowships: is it a problem?
13
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Am J Roentgenol 1998;170(3):577-81. 45. Misakian AL, Bero LA. Publication bias and research on
27. Dale JA, Schmitt CM, Crosby LA. Misrepresentation of passive smoking: comparison of published and
research criteria by orthopaedic residency applicants. J unpublished studies. JAMA 1998;280(3):250-3.
Bone Joint Surg 1999;81(12):1679-81. 46. McNutt RA, Evans AT, Fletcher RH, Fletcher SW. The
28. Tangney JP. Fraud will out—or will not? New Scientist effects of blinding on the quality of peer review. JAMA
1987;115(August 6):62-3. 1990;263(10, March 9):1371-76.
29. Korenman SG, Berk R, Wenger NS, Lew V. Evaluation 47. Black N, van Rooyen S, Godlee F, Smith R, Evans S.
of the research norms of scientists and administrators What makes a good reviewer and a good review for a
responsible for academic research integrity. JAMA general medical journal? JAMA 1998;280(3):231-3.
1998;279(1):41-7. 48. Callaham ML, Baxt WG, Waeckerle JF, Wears RL.
30. Committee on Science Engineering and Public Policy Reliability of editors’ subjective quality ratings of peer
(U.S.). Panel on Scientific Responsibility and the reviews of manuscripts. JAMA 1998;280(3):229-31.
Conduct of Research. Responsible science: Ensuring the 49. Cho MK, Justice AC, Winker MA, Berlin JA, Waeckerle
integrity of the research process. Washington, D.C.: JF, Callaham ML, et al. Masking author identity in peer
Nationl Academy Press; 1992. review: what factors influence masking success? JAMA
31. Rennie D, Flanagin A. Congress on Biomedical Peer 1998;280(3):243-5.
Review: history, ethics, and plans for the future 50. Godlee F, Gale CR, Martyn CN. Effect on the quality of
[editorial]. JAMA 1998;280(3):213. peer review of blinding reviewers and asking them to
32. Pitkin RM, Branagan MA. Can the accuracy of abstracts sign their reports: a randomized controlled trial. JAMA
be improved by providing specific instructions? A 1998;280(3):237-40.
randomized controlled trial [see comments]. JAMA 51. Justice AC, Cho MK, Winker MA, Berlin JA, Rennie D.
1998;280(3):267-9. Does masking author identity improve peer review
33. Gardner MJ. An exploratory study of statistical quality? A randomized controlled trial. JAMA
assessment of papers published in the British Medical 1998;280(3):240-2.
Journal. JAMA 1990;263(March 10):1355-57. 52. van Rooyen S, Godlee F, Evans S, Smith R, Black N.
34. Eichorn P, Yankauer A. Do authors check their Effect of blinding and unmasking on the quality of peer
references? A survey of accuracy of references in three review: a randomized trial. JAMA 1998;280(3):234-7.
public health journals. Am J Public Health 53. van Rooyen S, Black N, Godlee F. Development of the
1987;77(8):1011-2. review quality instrument (RQI) for assessing peer
35. Evans JT, Nadjari HI, Burchell SA. Quotational and reviews of manuscripts. J Clin Epidemio
reference accuracy in surgical journals. JAMA 1999;52(7):625-9.
1990;263(10):1353-4. 54. Armstrong JS. Peer review for journals: Evidence on
36. Garfunkel JM, Ulshen MH, Hamrick HJ, Lawson EE. quality control, fairness, and innovation. Sci Eng Ethics
Problems identified by secondary review of accepted 1997;3(1):63-84.
manuscripts. JAMA 1990;263(10):1369-71. 55. Weber EJ, Callaham ML, Wears RL, Barton C, Young
37. Stewart WW, Feder N. The integrity of the scientific G. Unpublished research from a medical specialty
literature. Nature 1987;325:207-14. meeting: why investigators fail to publish. JAMA
38. Garfunkel JM, Ulshen MH, Hamrick HJ, Lawson EE. 1998;280(3):257-9.
Effect of institutional prestige on reviewers’ 56. Fletcher RH, Fletcher SW. Evidence for the
recommendations and editorial decisions. JAMA effectiveness of peer review. Sci Eng Ethics
1994;272(2):137-8. 1997;3(1):35-50.
39. Joyce J, Rabe-Hesketh S, Wessely S. Reviewing the 57. Friedman PJ. Correcting the literature following
reviews: the example of chronic fatigue syndrome. fraudulent publication. JAMA 1990;263(10, March
JAMA 1998;280(3):264-6. 9):1416-19.
40. Link AM. US and non-US submissions: an analysis of 58. Garfield E, Welljams-Dorof A. The impact of fraudulent
reviewer bias. JAMA 1998;280(3):246-7. research on the scientific literature. JAMA
41. Jadad AR, Cook DJ, Jones A, Klassen TP, Tugwell P, 1990;263(10):1424-27.
Moher M, et al. Methodology and reports of systematic 59. Pfeifer MP, Snodgrass GL. The continued use of
reviews and meta-analyses: a comparison of Cochrane retracted, invalid scientific literature. JAMA
reviews with articles published in paper-based journals. 1990;263(10):1420-3.
JAMA 1998;280(3):278-80. 60. Budd JM, Sievert M, Schultz TR. Phenomena of
42. Dickersin K, Fredman L, Flegal KM, Scott JD, Crawley retraction: reasons for retraction and citations to the
B. Is there a sex bias in choosing editors? Epidemiology publications. JAMA 1998;280(3):296-7.
journals as an example. JAMA 1998;280(3):260-4. 61. Parrish DM. Scientific misconduct and correcting the
43. Dickersin K. The existence of publication bias and risk scientific literature. Acad Med 1999;74(3):221-30.
factors for its occurrence. JAMA 1990;263(10):1385-9. 62. Snodgrass GL, Pfeifer MP. The characteristics of
44. Callaham ML, Wears RL, Weber EJ, Barton C, Young medical retraction notices. Bull Med Libr Assoc
G. Positive-outcome bias and other limitations in the 1992;80(4):328-34.
outcome of research abstracts submitted to a scientific 63. Duggar DC, Christopher KA, Tucker BE, Jones DA,
meeting. JAMA 1998;280(3):254-7.

14
––––––––––––––––––––––––––––––––––––––––––––––––––––– Steneck, Assessing the Integrity of Publicly Funded Research
Watson M, Puckett M, et al. Promoting an awareness of study of confirmatory bias in the peer review system.
retractions: the Louisiana State University Medical Cognitive Therapy and Research 1977;1:161-75.
Center in Shreveport experience. Medical Reference 82. Dickersin K, Min Y-I, Meinert CL. Factors influencing
Services Quarterly 1995;14(1):17-32. publication of research results. Follow-up of
64. Editors ICoMJ. Uniform requirements for manuscripts applications submitted to two institutional review
submitted to biomedical journals. JAMA boards. JAMA 1992;267(3):374-8.
1993;269:2282-6. 83. Campbell TID. Understanding the potential for
65. Flanagin A, Carey LA, Fontanarosa PB, Phillips SG, misconduct in university-industry relationships: An
Pace BP, Lundberg GD, et al. Prevalence of articles with empirical study. In: Braxton JM, editor. Perspectives on
honorary authors and ghost authors in peer-reviewed scholarly misconduct in the sciences. Columbus, OH:
medical journals. JAMA 1998;280(3):222-4. Ohio State University Press; 1999. p. 259-82.
66. Hoen WP, Walvoort HC, Overbeke AJ. What are the 84. Boyd EA, Bero LA. Assessing Faculty Financial
factors determining authorship and the order of the Relationships With Industry : A Case Study. JAMA
authors’ names? A study among authors of the 2000;284:2209-14.
Nederlands Tijdschrift voor Geneeskunde (Dutch 85. Cho MK, Shohara R, Schissel A, Rennie D. Policies on
Journal of Medicine). JAMA 1998;280(3):217-8. faculty conflicts of interest at US universities. JAMA
67. Shapiro DW, Wenger NS, Shapiro MF. The 2000;284:2203-8.
contributions of authors to multiauthored biomedical 86. Health NIo. Reminder and update: Requirement for
research papers. JAMA 1994;271(6):438-42. instruction in the responsible conduct of research in
68. Tarnow E. The authorship list in science: Junior national Research Service Award institutional training
physicists’ perceptions of who appears and why. Sci Eng grants. NIH Guide for Grants and Contracts 1992;21:2-
Ethics 1999;5(1):73-88. 3.
69. Drenth JP. Multiple authorship: the contribution of 87. Mastroianni A, Kahn JP. Encouraging accountability in
senior authors. JAMA 1998;280(3):219-21. research: A pilot assessment of training efforts.
70. Wilcox LJ. Authorship: the coin of the realm, the source Accountability in Research 1999;7:85-100.
of complaints. JAMA 1998;280(3):216-7. 88. Brown S, Kalichman MW. Effects of training in the
71. Health NIo. Guidelines for the Conduct of Research in responsible conduct of research: A survey of graduate
the Intramural Research Program at NIH. [Bethesda, students in experimental sciences. Sci Eng Ethics
MD]: [National Institutes of Health]; 1997. 1998;4(4):487-98.
72. Waldron T. Is duplicate publishing on the increase? Br 89. Elliott D, Stern JE. Evaluating teaching and students’
Med J 1992;304:1029. learning of academic research ethics. Sci Eng Ethics
73. Barnard H, Overbeke AJ. [Duplicate publication of 1996;2(3):345-66.
original manuscripts in and from the Nederlands 90. Self DJ, Wolinsky FD, Baldwin DC, Jr. The effect of
Tijdschrift voor Geneeskunde]. Ned Tijdschr Geneeskd teaching medical ethics on medical students’ moral
1993;137(12):593-7. reasoning. Acad Med 1989;64:755-9.
74. Koene HR, Overbeke AJ. [The ultimate fate of articles 91. Baldwin DC, Jr., Daugherty SR, Self DJ. Changes in
rejected for publication in the Nederlands Tijdschrift moral reasoning during medical school. Acad Med
voor Geneeskunde]. Ned Tijdschr Geneeskd 1991;69(9 Suppl):S1-3.
1994;138(49):2443-6. 92. Self DJ, Schrader DE, Baldwin DC, Jr., Wolinsky FD. A
75. Blancett SS, Flanagin A, Young RK. Duplicate pilot study of the relationship of medical education and
publication in the nursing literature. Image: The Journal moral development. Acad Med 1991;66(10):629.
of Nursing Scholarship 1995;27(1):51-6. 93. Self DJ, Schrader DE, Baldwin DC, Jr., Root SK,
76. Bloemenkamp DGM, Walvoort HC, Hart W, Overbeke Wolinsky FD, Shadduck JA. Study of the influence of
AJPM. [Duplicate publication of articles in the Dutch veterinary medical education on the moral development
Journal of Medicine in 1996]. Ned Tijdschr Geneeskd of veterinary students. J Am Vet Med
1999;143(43):2150-3. Assoc1991;198(5):782-7.
77. Tramer MR, Reynolds DJ, Moore RA, McQuay HJ. 94. Self DJ, Baldwin DC, Jr., Wolinsky FD. Evaluation of
Impact of covert duplicate publication on meta-analysis: teaching medical ethics by an assessment of moral
a case study. Br Med J 1997;315(7109):635-40. reasoning. Med Educ 1992;26(3):178-84.
78. Jefferson T. Redundant publication in biomedical 95. Self DJ, Schrader DE, Baldwin DC, Jr., Wolinsky FD.
sciences: Scientific misconduct or necessity? Sci Eng The moral development of medical students: a pilot
Ethics 1998;4(2):135-40. study of the possible influence of medical education.
79. Merton RK. A note on science and democrary. Journal Med Educ 1993;27(1):26-34.
of Legal and Political Sociology 1942;1(1-2):115-26. 96. Bebeau MJ, Tohma SJ. The impact of a dental ethics
80. Jasanoff SJ. Is science socially constructed: And can it curriculum on moral reasoning. J Dent Educ
still inform public policy? Sci Eng Ethics 1994;58(9):684-92.
1996;2(3):263-76. 97. Self DJ, Baldwin DC, Jr., Wolinsky FD. Further
81. Mahoney MJ. Publication prejudices: An experimental exploration of the relationship between medical
education and moral development. Camb Q Healthc

15
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Ethics 1996;5(3):444-9.
98. Anderson MS, Louis KS, Earle J. Disciplinary and
departmental effects on observations of faculty and
graduate student misconduct. In: Braxton JM, editor.
Perspectives on scholarly misconduct in the sciences.
Columbus, OH: Ohio State University Press; 1999. p.
213-35.
99. Clarke M, Chalmers I. Discussion sections in reports of
controlled trials published in general medical journals:
islands in search of continents? JAMA 1998;280(3):280-
2.
100. Junker CA. Adherence to published standards of
reporting: a comparison of placebo-controlled trials
published in English or German. JAMA
1998;280(3):247-9.
101. Begg C, Cho MK, Eastwood S, Horton R, Moher D,
Oking I, et al. Improving the quality of reporting of
randomized controlled tirals: the CONSORT statement.
JAMA 1996;276:637-9.
102. Pitkin RM, Branagan MA, Burmeister LF. Accuracy of
data in abstracts of published research articles. JAMA
1999;281:1129-30.
103. Pitkin RM, Branagan MA, Burmeister LF.
Effectiveness of a journal intervention to improve
abstract quality. JAMA 2000;283(4):481.
104. Jefferson T, Smith R, Yee Y, Drummond M, Pratt M,
Gale R. Evaluating the BMJ guidelines for economic
submissions: prospective audit of economic submissions
to BMJ and The Lancet. JAMA 1998;280(3):275-7.
105. Lock S, Wells F. Fraud and misconduct in medical
research. London: BMA Publishing Group; 1993.
106. Braxton JM, ed. Perspectives on Scholarly Misconduct
in the Sciences. Columbus, OH: Ohio State University
Press; 1999.

16
I. Norms and Environmental Issues

1. Students and Mentors

What Would Get You in Trouble: Doctoral Students’ Conceptions of Science
and Its Norms
Melissa S. Anderson, Educational Policy and Administration, University of Minnesota, USA

Key Words: Graduate education, Graduate students, Norms

Undergirding the academic enterprise is a web of assumptions about how the members of the
academic community should conduct their professional lives. These assumptions are expressed in
ways ranging from the most explicit directives (legal, institutional, contractual) to the implicit, taken-
for-granted understandings that facilitate everyday interactions among members of the profession.
They constitute the normative underpinnings of the academic profession.
Braxton and Bayer define norms as “shared beliefs within a particular social or professional
group about behavior expected or desired in a given situation or circumstance” (1). In the academic
context, the four norms that Robert Merton (2) identified in his 1942 analysis–universalism,
communality [to use Barber’s (3) term], disinterestedness, and organized skepticism–have framed
much of the subsequent research. They figured prominently in Zuckerman’s seminal analyses of the
social system of science (4, 5). They are also reflected in Mitroff’s (6) “counternorms”, and they
together capture most of the considerable literature that Braxton (7) compiled on the subject of norms.
Others, however, have argued for a more complex understanding of norms. Mulkay, for example,
has claimed that norms are best understood as ideologies or “evaluative repertoires” (8). That is,
norms constitute a kind of standardized narrative that academics use to describe and evaluate
behavior and to prescribe responses to certain behaviors (8). Ajzen and Fishbein have described the
significance of “subjective norms” that reflect what others, who are important to an individual, think
he or she should do (9). From this perspective, neither an abstract normative system or an
individual’s own internalized norms are as important as the individual’s understanding of others’
expectations. Finally, Braxton and Bayer have demonstrated how a combination of inductive and
survey-based strategies could uncover a complex set of norms in collegiate teaching (1).
The present study takes a different approach to the norms of the academic profession, with
corresponding implications for the design of the study. First, it emphasizes the implicit over the
explicit, on the assumption that implicit norms can be particularly powerful in shaping behavior. This
study therefore relies on narrative descriptions of norms, instead of on a particular formulation of the
normative structure of academia. It is rooted in the proposition that more attention needs to be paid to
understanding science and its ethical aspects from the “inside out,” that is through the experiences of
scientists themselves (10-12). It therefore responds to Braxton’s call for study of norms “expressed in
the words of the respondents rather than in a priori definitions of possible norms” (7).
Second, it assumes that norms of a group are particularly salient to newcomers during a
socialization period (13). The data for this study accordingly come from first-year doctoral students,
who are encountering professional norms in intensive ways. Their experiences are likely to produce

Corresponding author: Melissa S. Anderson, Educational Policy and Administration, 330 Wulling Hall, University of
Minnesota, Minneapolis, MN, 55455, 612-624-5717 (voice), 612-624-3377 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
“contrast” in the gestalt sense through the process troubled by dissonance between behaviors and
of “sense-making”, which highlights the norms, recognizing the inevitable roles played by
normative insights they acquire (14). mistakes, errors of fact and of judgment, and
Third, the study assumes no necessary match mid-course corrections.
among students’ understanding of the broad
norms of the academic profession, the norms that Results
they have internalized and view as most salient, Students’ conceptions of norms that underlie
and the behavior of professional colleagues. This their work are presented here in terms of the
study therefore explores levels of consonance three contrasts identified above. First, students’
and dissonance that students perceive among conceptions of general academic norms are
these three phenomena. described in light of the behavior of their
Fourth, this study relies on Durkheim’s colleagues. Then the norms to which they
useful proposition that norms are recognized subscribe are seen in contrast, again, to
when they are violated (15). The questions used colleagues’ behavior. Finally, what they
in this study to elicit students’ views of norms, understand to be academic norms are contrasted
therefore, ask students to contrast their views of to their own normative orientations.
general academic norms, as well as the norms to Correspondence between academic norms
which they subscribe, against the behavior of and behavior. The first comparison investigated
their colleagues. is between students’ conceptions of the norms of
their fields and the behaviors of those around
Methods them. The interview question was, “Do you see
These parameters gave shape to the current study, any conflicts between what people think or say
which is part of a broader project on doctoral you should do and the way work is actually
education, the Academic Life Project, funded by done?”
the National Science Foundation (Grant number Approximately two-thirds of those
9408S08622). Participants for the current interviewed saw no conflict between prescribed
analysis were 30 first-year doctoral students in and actual behavior among their colleagues.
seven science and social science disciplines at a Most saw no disjuncture; a few were more
major research university. (The project will definite: “No, I mean, an emphatic no with the
eventually involve over 100 interviewees and faculty,” and, “They’re pretty straightforward,
will be longitudinal.) Semi-structured interviews and they’ll pretty much hold true to their word.”
of approximately a half-hour yielded narrative Two students noted that, while they were not
data on norms and related topics. aware of conflict between norms and action, they
A series of questions in the interviews asked did not really know enough about what people
students to consider and comment on were doing in the department to comment
relationships between academic norms and generally about people’s behavior; as one put it,
behavior (Do you see any conflicts between what “I’m not privy to a lot of the goings on of the
people think or say you should do and the way department.”
work is actually done?), between their own Five students noted particular areas of
perspectives and behavior (Do you see people disjuncture between norms and behavior. One
around here acting contrary to your advice [to mentioned safety rules:
doctoral students on how to avoid serious We all have to go to this safety training before
mistakes]?), and between their own normative we are allowed to go in the lab. It’s just kind
perspectives and academic norms (Are there any of a refresher course every year. And then ...
ideas or rules about how you should do your they always say practically nothing is supposed
work that you don’t agree with?). These to go down the drain. And sometimes stuff
does. But we’re not even supposed to put things
questions highlighted students’ understandings of like ... a simple rinsing agent down the drain ...
academic research as a social enterprise whose but it happens all the time.
membership they are entering. Those who This student went on to affirm the importance of
articulated a more complex normative the safety rules for two reasons: first, that safety
perspective showed greater awareness of the supports proper procedures (“if you don’t do it
social aspects of the scientific enterprise and a right, it doesn’t work”), and second, not
more constructivist approach to knowledge following these rules is dangerous (“if you don’t
development in the sciences. They were also less follow the rules in terms of safety, in terms of
20
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Anderson, What Would Get You in Trouble
correct procedure, usually that means that the in the interviews is between the norms to which
chemist should not work in the lab”). students themselves subscribe and the behavior
A second point of conflict observed by a of their colleagues. Here the question is whether
student is in admissions procedures for the or not students see people around them acting
graduate students in the department. From the contrary to the way the students think they
vantage point of a place on the departmental should act. Employing Durkheim’s view that
graduate admissions committee, the student saw norms are best recognized when violated, the
that, though the department touts a highly interview protocol invited students to consider
selective admissions policy, the process is what they would advise incoming doctoral
influenced in political ways by influential students to do to stay out of trouble in their work
individuals on the faculty. The result is that the (15). Responses demonstrate students’
department admits less-qualified people than its personally held beliefs about how first year
policy would suggest. students should act, identified here as subscribed
The third area of dissonance between norms. Students were then asked, as follow-up
prescribed and enacted behaviors is in research. questions, “Do you see people around here acting
One psychology student focused on experiments: contrary to your own advice? What are they
We talk a lot about being a very experimental doing?”
field and it’s all about experiments, but it’s so Responses to these questions fall into three
difficult to run experiments now with getting general categories: tasks, relationships, and
through the IRB [Institutional Review Board] ethics. Most of the responses addressed the work
and getting subjects.... [I]t’s so much easier to of graduate students. Several talked about the
pass out some sort of survey or some sort of
questionnaire. And so we talk about the
need for students to take responsibility for their
experiment and how wonderful it is, and then own work and progress. As one put it, “I mean,
we don’t do it. in our department, it’s a problem both with the
Two other students also mentioned research, but students not taking the initiative to getting all of
in a different way. They clearly understood the their requirements and prelims done and also,
faculty’s focus on research, but they did not see with our department, no one says anything if it
faculty providing enough support to students to takes you longer.” Others disapproved of student
get them started on their own research. As one colleagues’ not getting their work done, taking
put it, “I think [it’s] the absence of direction too much time to get through their work, or
which is noticeable, which stands out. And I abandoning the work altogether. All of these
think some students have felt ... you know, students clearly demonstrated a strong
they’re sort of cast adrift, in some sense, and left commitment to hard work and a sense that some
to figure everything out for themselves.” The others around them acted contrary to this
other student described her frustration with the subscribed norm.
research imperative in light of the same kind of Not only do students believe in getting the
lack of direction: work done, but several mentioned the need to do
There almost seems like there’s kind of pressure independent work. One science student
or an expected norm within the department complained,
itself that we get involved with research. Yet, I think one of the biggest mistakes that they
in our specific discipline, in our area, there could make is to do something that is not
hasn’t been very much guidance or, you know, independent. I see a lot of people that are
pressure to do that.... I have met with my working with their advisors and really, ... I don’t
advisor twice on my own volition — and going know the best way to describe this without
to see her and saying, “Okay. Maybe it’s time sounding mean, but they just have no interest
for me to get involved in research,” and each of their own. They are just a, like a little off-
time she has not had a specific project that shoot of their advisor, like a little worker....
really had any place for me to start.... And I They’re not independent at all.... You know,
just kind of walked away from it feeling like, what they do is what their advisor says, and I
just thinking that she had just so much going think that’s a really big mistake, because one
on already — and really, you know, like almost day you can look back and be, like, “Oh. This
I kind of felt like I would be a burden to get isn’t what I wanted to do at all, and if I had the
involved at that point. choice I would have done it completely
Correspondence between subscribed norms differently.”
and behavior. The second comparison addressed Taking the initiative for an independent stream of
21
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
inquiry is a step beyond responsibility for getting I guess, just like, you have to do things for
work done, a step that some, but not all, first-year people so they’ll do things for you later. I guess
graduate students take. that doesn’t even sound that bad. But more
One student’s story about a graduate-student like — I can’t think of a particular example —
peer illustrates her struggle with maintaining her but just basically doing things that you don’t
want, because you know later it’ll get you
independence in inquiry. The peer in question is something you do want.
someone she respects. Not only are students aware of the work
But the problem is, he comes from a different
undergraduate background, not an American
imperative, but they are also aware of the need
system. He’s from a different country, where for others to know that they subscribe to it. As
being the best in the class was very much the quotes illustrate, the norm bears both sanction
recognized and very much rewarded, and so and reward. This norm illustrates students’
he was the best in his class. And so he came movement toward full acceptance into the
here.... Everyone has been asking him for help, academic social world.
and so he would do all of his work way in The third contrast between behavior and
advance — which was commendable — but students’ own normative orientations was in the
then he would — instead of working and taking area of ethics. Those who mentioned ethics said
other people’s suggestions and trying to that they had seen no instances of people acting
integrate everything when we were working on contrary to what they themselves understood to
problem sets — he would be, like, “This is
right. I have the answer.” And usually he did.
be appropriate behavior. One said, “I’ve never
Usually he was right. But it was annoying to seen anyone falsifying data, which is very, very
work with him.... There were times where even good. And I believe that we don’t have the
though I knew I would probably get a better second problem, fishing for data. At least in my
grade if I worked with him, because he would group, we don’t have that.” Another noted, “I
have the answers, I wouldn’t want to do it. And haven’t seen, I haven’t heard of anybody lying
also, you don’t want the answers given to you. about stuff or trying to falsify results.” This
Comments about relationships comprise the science student went on to describe how
next category of responses about the contrast important it is for students to acknowledge
between subscribed norms and behavior. mistakes, so that they are not interpreted as more
Students demonstrate clear ideas about how serious offenses: “Everybody makes mistakes....
people should behave toward each other in the Everyone’s pretty understanding of when your
graduate school setting. Some mentioned the experiments don’t work or when you did a stupid
importance of having an advisor with whom the mistake or whatever.”
student can work. They described examples of The normative understandings that the
advisors who were not supportive of their doctoral students reveal through their comments
students. This behavior that ran contrary to their on the contrast between what peers should do and
beliefs about how advisors are to act met with what they are actually doing thus center largely
very strong negative reactions. on their work and their relationships with
Other respondents showed a keen sense of colleagues. That is, they appear attuned to both
the importance of making a good impression and functional and social norms of academic life.
expressed dismay that some of their peers did not The next step is to contrast their own normative
appear to understand this point. A science orientations to what they perceive to be the
student said, general norms of their fields.
Contrast between academic norms and
I know there’s some people who, whenever
there was an exam, they just didn’t go into the subscribed norms. Students’ perceptions of
lab all the time, and I don’t think it left a good prevalent academic norms may not match their
impression on some people who were working own ideas about how they should conduct
in the lab, working around them.... So if you themselves in the academic world. As both
don’t seem very serious about your lab work, academic norms and subscribed norms can be
then they — someday when you have to go to brought into focus by contrasting them against
them for advice or something — they’re not behavior, so they can be clarified by comparing
necessarily drawn to give you as much time them to each other. The relevant question on the
and make as much of a serious effort. interview protocol was, “Are there any ideas or
Another student described impression- rules about how you should do your work that
management in blunt terms as a quid pro quo: you don’t agree with?”
22
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Anderson, What Would Get You in Trouble
The task-related points of disjuncture fell Beyond students’ attention to task-related
generally in the category of competition and its disjunctures between academic and espoused
attendant work pressures. A student in a social norms, the most striking pattern in students’
science department commented, responses is their uncertainty about academic
Everyone’s competing for jobs in academic norms in general. Most of them are keenly aware
environments primarily.... And I guess what that that norms vary by discipline or even from one
means for many students is they have to adapt research group to another. For example, one
to a competitive type of atmosphere and in noted, “Everyone has such different views about
some cases be more competitive than they how to do things.” Another put it this way:
would like to be in order to further their goals
further on. And I think that might be
“Each professor sort of has their own research
disheartening for some students.... And I think policy. And that’s academia. They have the
all of the students ... try to be good-natured freedom to make up the rules of their group,
about the entire thing, but I think the pressure within certain bigger boundaries that the school
of continuing to get as many publications as sets.” Yet another respondent said, “I don’t think
you can is the reality that dawns on a lot of there are very many rules about how we should
students — something they didn’t anticipate, conduct our research, other than the whole basic
necessarily, early on. ‘Be ethical and stuff.’ I don’t observe very many
Another student talked about competitive rules about how we should conduct the research.”
pressures to publish in terms of “the whole This student went on to mention that she might
production thing” and the “assembly line change her mind as she got further into her
production attitude.” research, when she would have to remember all
Several students complained about the work the rules about where to put commas — thereby
loads they bear in terms of the mismatch between illustrating just how far she had to stretch to
their professors’ views on how much work they think of general norms of the field.
should do and their own. A science student Perhaps some of the uncertainty that students
talked about peers who never take time off and expressed about academic norms is related to the
“work themselves to death” to live up to what ways in which such norms are communicated.
they perceive as the standards of work in the The student quoted above who mentioned each
field; the student said he would never do that. professor having his or her own research policy
Another commented on prevalent norms for the went on to say, “Ideally, it should be talked about
quality of a dissertation. In this students’ as a research group as a whole, but it seems to
relatively new field in science, it was generally me that a lot of stuff is just sort of telephone,
expected, 10 or 20 years ago, that each where one person tells another person, and that
dissertation would open up a completely new person tells the next person.” Another talked
field of inquiry; now, the expansion of the about his reluctance to ask people how things
discipline and the far greater competition due to a should be done in the lab:
more crowded field make it much harder to have The approach towards how you learn your way
such an impact through doctoral work. The around the lab is you just go in there and you
student noted, though, that normative do it. As far as being taught or having anyone
understandings in the field had not changed in specifically to show you around, you really
response. don’t, because everyone in there is really, really
Another point of contrast related to busy, because they are doing research. And
competition is the matter of independent work. they don’t want to take time out of their
research to show you how to work [a machine],
Several students mentioned that at least some of because it’s such a simple thing to them, and
their professors require independent, as opposed they get really frustrated and impatient with
to collaborative, work on assignments in someone who is just learning how to use it.
graduate courses. Many of the students were And so, generally you just have to go in there
previously socialized to collaborative norms, and and learn on your own.... I almost felt afraid to
they found the professors’ insistence on go to other people in the group with my stuff,
individual work counterproductive. Here because I don’t want to waste their time and I
students’ normative orientations run counter to don’t want to feel stupid either.
the academy’s norms of rewarding people on the Of course, some students were unable to identify
basis of individual achievement and independent any dissonance between the norms to which they
contributions. subscribe and the more general academic norms
23
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
as they see them. One person wryly commented furthermore suggests ways in which those norms
on the thoroughness of his own socialization to can be communicated within and beyond the
the general normative structure of the field: scientific community (17, 18). The doctoral
“Maybe I’ve been just so well trained that I don’t students interviewed reveal the norms of science
know anything anymore.” as they understand them, during a period when
The results in this section show, as did the they are intensely and reflectively face-to-face
earlier results, that students’ normative with the way science works. They are the future
conceptions are dominated by functional or task- membership of the scientific community, but they
related norms. They also show a general are also current participants in the enterprise,
awareness among students of social norms, struggling with their own ideas of how they
though their conceptions of norms for should behave as scientists.
interpersonal relations are not as fully developed The results of the interviews demonstrate
as their views on functional norms. intriguing patterns of dissonance among the three
phenomena examined. The interview responses
Discussion show that students’ normative conceptualizations
The findings presented here contribute to our are dominated by functional (task-related) norms,
understanding of doctoral students’ initial as we might expect from earlier work on
normative orientations. Students’ conceptions of anticipatory socialization that emphasizes
normative imperatives are relevant to policy survival in the graduate or professional-school
initiatives that are currently receiving a great deal setting (16). Augmenting the functionalist
of attention. The federal Office of Research perspective, however, are emergent
Integrity recently announced a major new conceptualizations of social and ethical norms.
initiative that will focus on the promotion of the The inchoate nature of first-year students’
responsible conduct of research. The American personal normative orientations suggests that
Educational Research Association is currently approaches to socialization of doctoral students
preparing to publish a book that will direct to academic life, particularly in the areas of
attention to the AERA Code of Ethics and its use. ethics and related issues, may overestimate the
Dozens of other academic associations are extent of students’ understanding of the academic
writing or revising their codes of ethics, and system, the nature of research, and the place of
virtually every major research university has individual academics in the broader context of
addressed its institutional policies on ethics and research. Students interviewed here showed very
misconduct in the past five years. The federal little awareness of their disciplines, beyond their
government is seeking to expand its requirements own work, or of the higher education system,
for formal training in ethics beyond those for beyond their own departments. The imperatives
trainees covered by National Institutes of Health they identified have to do generally with the
funding. Most of the attention to expanded work at hand and the people with whom they
training in ethics and related issues focuses on interact.
graduate students and other newcomers to the Socialization to the field and to the
academic profession. normative bases of research in a discipline should
Continued self-regulation by the scientific be grounded in the academic world with which
community depends on the ongoing renewal of these students are familiar, while at the same
normative conceptualizations that, through their time introduce them to the broader academic
generational evolution, continue to reflect the environment. The theme of individual,
expectations of society for science. Most of the independent work that runs through these
emerging initiatives are driven, however, by a interviews suggests that students might not be
sense of urgency or by federal regulations and subject to as much osmotic group socialization as
directives, without attention to doctoral students’ many faculty assume. It is also clear that the
understanding of science, academic life, and the channels by which socialization to the normative
norms of their disciplines. Neither do they aspects of academic life are communicated are
reflect ways in which newcomers interact with primarily informal. Calls for more formal, more
and shape the normative bases of their fields deliberate approaches to normative socialization
(16). find support in the vagueness with which
This study serves as a window onto the students conceptualize the norms that underlie
normative assumptions of science, but it academic research.
24
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Anderson, What Would Get You in Trouble

Aknowledgements 17. Lapidus JB., Mishkin B. Values and ethics in the

The author gratefully acknowledges the research graduate education of scientists. In: Nerad M, June R,
assistance of Janet M. Holdsworth and Jennifer Miller DS, editors. Graduate education in the United
States. New York: Garland; 1997.
Milleville. 18. Bird SJ. Including ethics in graduate education in
scientific research. In: Braxton JM, editor. Perspectives
Bibliography on scholarly misconduct in the sciences, Columbus
1. Braxton JM, Bayer AE. Faculty misconduct in collegiate (OH): Ohio State University Press; 1999.
teaching. Baltimore: The Johns Hopkins University
Press; 1999.
2. Merton RK. A note on science and democracy. J of
Legal and Political Sociology 1942; 1(1-2):115-126.
3. Barber B. Science and the social order. New York: Free
Press; 1952.
4. Zuckerman HE. Deviant behavior and social control in
science. In: Sagarin E, editor. Deviance and social
change. Beverly Hills (CA): Sage; 1977. p. 87-138.
5. Zuckerman HE. The sociology of science. In: Smelser
NJ, editor. Handbook of sociology. Newbury Park (CA):
Sage; 1988. p. 511-574.
6. Mitroff I. Norms and counter-norms in a select group of
the Apollo moon scientists: A case study of the
ambivalence of scientists. Amer Sociological Rev 1974;
39, 579-95.
7. Braxton JM. The normative structure of science: Social
control in the academic profession. In: Smart JC, editor,
Higher education: Handbook of theory and research, v.
2. New York: Agathon Press; 1986.
8. Mulkay, M. Norms and ideology in science. Soc Sci
Info 1976; 15(4-5): 637-56.
9. Ajzen I, Fishbein M. Understanding attitudes and
predicting social behavior. Englewood Cliffs (NJ):
Prentice-Hall; 1980.
10. Anderson MS. (1999). Uncovering the covert:
Research on academic misconduct. In: Braxton JM,
editor. Perspectives on scholarly misconduct in the
sciences. Columbus (OH): Ohio State University Press;
1999.
11. Knorr-Cetina KD, Mulkay M, editors. Science observed:
Perspectives on the social study of science. Beverly
Hills (CA): Sage; 1983.
12. Latour B, Woolgar S. Laboratory life: The construction
of scientific facts. Princeton, NJ: Princeton University
Press; 1986.
13. Wanous JP. Organizational entry: Recruitment,
selection, orientation, and socialization of newcomers,
2nd ed. Reading (MA): Addison-Wesley; 1992.
14. Louis MR. Surprise and sense making: What
newcomers experience in entering unfamiliar
organizational settings. Admin Sci Q 1980; 25(2):226-
251.
15. Durkheim, E. The elementary forms of religious life.
Fields KE, translator. New York: Free Press;
1995[1912].
16. Tierney WG., Rhoads RA. Enhancing promotion,
tenure, and beyond: Faculty socialization as a cultural
process. ASHE-ERIC Higher Education Report, no. 93-
6. Washington, D.C.: The George Washington
University; 1993.

25
Data Manipulation in the Undergraduate Laboratory: What are we
teaching?
Elizabeth W. Davidson, Department of Biology, Arizona State University, USA

Heather E. Cate, College of Liberal Arts and Sciences, Arizona State University, USA

Cecil M. Lewis, Jr., Department of Anthropology, University of New Mexico, USA

Melanie Hunter, Department of Biology, Arizona State University, USA

Keywords: Biology, Cheating, Chemistry, Motivation, Undergraduate research

Arizona State University (ASU) offers a senior-level course entitled “Professional Values in Science”
that addresses a number of topics concerning ethical conduct of research as well as ethical concerns at
the intersection of science and society. The course demands active participation by the students.
Several years ago, on his own initiative a student in the class developed a questionnaire that explored
data manipulation. As most of the students were undergraduates, the questionnaire focused upon
manipulation of data in undergraduate science laboratories. We were startled to discover that over
60% of the students openly admitted to manipulation of data in undergraduate laboratories. These
results led to the development of a more elaborate survey that has been administered to 7
undergraduate Biology and Chemistry courses, enrolling a total of over 700 students. The courses
include both major and nonmajor subjects, at both introductory and upper division level. Arizona
State University has approximately 50,000 students, including (in academic year 2000) ca. 1000
majors in Biology and 250 majors in Chemistry. In the fall semester, 2000, 3137 undergraduates are
enrolled in Biology courses, while 3355 undergraduates are enrolled in Chemistry courses.
Laboratories are therefore limited in available time, are generally supervised and graded by graduate
teaching assistants, and many, but not all, of these courses rely upon traditional laboratory exercises.

Methods:
The survey and instructions to students are presented in at the end of the paper. Students were
advised by the person administering the survey (who was not their course professor or teaching
assistant) that the results would be held anonymous and would not affect their grade. The courses
included Chemistry 115: Introductory, non-majors; Chemistry 335: Organic, non-majors; Biology
201: Anatomy and Physiology, non-majors; Biology 100: Introductory, non-majors; Biology 182:
Introductory, majors; Biology 193: Introductory, majors, critical thinking focus; Biology 385:
Invertebrate Zoology, majors. Seven hundred and two students participated. Institutional Human
Subjects committee approval was obtained. Data were analysed by Spearman correlation.

Corresponding author: Elizabeth W. Davidson, Department of Biology, Arizona State University, Box 871501, Tempe, AZ
85287-1501, 480-965-7560 (voice), 480-965-2519 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
100% 100%

Percent Data Manipulation in Other Courses

90% 90%
Percent Data Manipulation in This Course

80% 80%

70% 70%

60% 60%

50% 50%

40% 40% never

seldom
30% never 30% sometimes
seldom often
20% sometimes 20% almost always
often
10% almost always 10%

0% 0%
CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385 CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385

Course Prefix and Number Course Prefix and Number

Figure 1. Results of survey, Question 5, “Have you ever Figure 2. Results of survey, Question 10, “Have you ever
manipulated data or made up data in this course?” CHM manipulated or made up data in any other science course?”
115: Introductory, non-majors, N=86; CHM 335: Organic, CHM 115, N=87; CHM 335, N=52; BIO 201, N=27; BIO
non-majors, N=44; BIO 201: Anatomy and Physiology, non- 100, N=81; BIO 182, N=40; BIO 193, N=57; BIO 385,
majors, N=29; BIO 100: Introductory, non-majors, N=200; N=66. N= total number of responses to the specific
BIO 182: Introductory, majors, N=40; BIO 193: question.
Introductory, majors, critical thinking focus, N=57; BIO
385: Invertebrate Zoology, majors, N=64. N= total number always” was in BIO 193, however a large
of responses to the specific question. proportion of the remainder reported
manipulation “often” (Figure 1). Within the two
Results non-majors chemistry courses surveyed, less data
The key question in this survey was Question 5,
manipulation was found in CHM 115
“Have you ever manipulated data in this course?”
(Introductory) than in CHM 335 (Organic), and
As shown in Figure 1, between 40.4 and 75% of
indeed the highest overall reported manipulation
students in the surveyed course admitted to
(90.5% “almost always” or “often”) was reported
manipulating data “almost always,” and another
in Organic Chemistry. Conversations with
20-43.9% admitted to such manipulation “often.”
students in the Professional Values in Science
Students reporting data manipulation “seldom”
class and elsewhere confirmed that many have
represented less than 5% of those surveyed, and 100%
Percent Data Manipulation observed in this course

only one student out of over 500 who responded

90%
to this question replied “never.” Using
correlation analysis, we learned that admission of 80%
manipulation in the course surveyed was strongly 70%
correlated to admission of manipulation in other
60%
courses (Spearman Correlation Sig. (2 tailed)
0.355, significant at 0.01 level) (Figure 2). 50%
We asked whether data manipulation was 40%
related to the level (i.e. introductory vs.
30%
advanced) of the course, and whether the course never
seldom
was designed for majors or non-majors. No 20%
sometimes
significant difference was found between data 10% often
almost always
manipulation in Introductory Biology BIO 100
0%
(non-majors) and BIO 182 (majors) or between
CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385
these lower division courses and an upper Course Prefix and Number
division course, BIO 385 (Invertebrate Zoology, Figure 3. Results of survey, Question 7, “Have you ever
majors). We compared responses from BIO 182, observed anyone else manipulate or make up data in this
a traditional introductory course, to BIO 193, an course?” CHM 115, N=91; CHM 335, N=67; BIO 201,
introductory majors course with emphasis on N=28; BIO 100, N=237; BIO 182, N=40; BIO 193, N=56;
critical thinking. The smallest percentage of BIO 385, N=66. N= total number of responses to the
students reporting data manipulation “almost specific question.
28
–––––––––––––––––––––––––––––––––––––––––––––– Davidson, et al., Data Manipulation in the Undergraduate Laboratory
100% 100%

Percent Data Manipulation for Better Grade, This Course

90% 90%

80% 80%
Percent Data Manupulation

70% 70%

60% 60%

50% 50%

40% 40%
never
seldom
30% 30% never
sometimes
seldom
often
20% 20% sometimes
almost always
often
10% 10% almost always

0% 0%
CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385 CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385
Course Prefix and Number Course Prefix and Number
Figure 4. Results of survey, Question 14, “Have you ever Figure 5. Results of survey, Question 6, “If you have ever
observed anyone manipulate or make up data in any science manipulated data or made up data, was it motivated by the
course?” CHM 115, N=94; CHM 335, N=70; BIO 201, thought of a better grade?” CHM 115, N=69; CHM 335,
N=30; BIO 100, N=96; BIO 182, N=39; BIO 193, N=55; N=41; BIO 201, N=17; BIO 100, N=246; BIO 182, N=31;
BIO 385, N=66. N= total number of responses to the BIO 193, N=55; BIO 385, N=53. N= total number of
specific question. responses to the specific question.
manipulated data for Chemistry laboratory not shown; Spearman correlation 0.368,
reports, particularly in Organic. Little difference significant at 0.01 level). Finally we surveyed
in data manipulation (Question 5) was found student preferences for type of laboratory
when analyzed by academic year or by gender. experiments (Question 17). In all seven courses
Two other key questions were 7 and 14, combined, only 1.7% of students preferred lab
which asked whether the student had observed experiments which place more emphasis on
others manipulating data. The results from these results, whereas 53.5% preferred more emphasis
questions were less consistent than responses to be placed upon processes, and 44.7% preferred
about the students own data manipulation. Two a balanced combination of both techniques
courses (CHM 115 and BIO 201) received an (N=503).
“almost always” response rate of 100%,
100%
whereas in other courses a much smaller
Percent Data Manipulation because Grade

proportion of students responded “almost 90%

always” (Figures 3, 4). 80%
Depends on Experimental Results

We investigated motivation for data

70%
manipulation with questions 6 and 11, which
asked whether the students manipulated data 60%
in order to get a better grade. Up to 100% of 50%
students in some courses replied that
manipulation was almost always performed to 40%

obtain a better grade (Spearman Correlation 30% never

Sig. (2-tailed) 0.265, significant at 0.01 level) seldom
20%
(Figure 5; data from Question 11 not shown). sometimes
often
When asked whether this was because they 10%
almost always
felt that their grade depended heavily upon the 0%
experimental results (Questions 8 and 15), less CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385
than half of students felt that their grade in the Course Prefix and Number
current course depended on experimental Figure 6. Results of survey, Question 8, “How often have
results “almost always”, and from 3.0 to you felt as though your grade in this course depended
13.6% of the students replied to Question 8 heavily on your experimental results?” CHM 115, N=102;
that their grade “seldom” depended on results CHM 335, N=66; BIO 201, N=35; BIO 100, N=218; BIO
(Figure 6, data from Question 15 not shown; 182, N=40; BIO 193, N=58; BIO 385, N=66. N= total
Spearman (Figure 7, data from Question 16 number of responses to the specific question.
29
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Class Places too much Emphasis on Experimental Results
uniformly high” (1). In surveys taken from 1970
100%
to present, from 42% to over 90% of students
90% reported cheating on tests by self or others
(reviewed by Cizek, (1)). Out of 6000 students
80%
in 31 universities surveyed by Meade, 67% of
70% science majors reported cheating on tests (2).
60%
Most surveys of college test cheating ask only
whether the student has ever cheated. Our survey
50% expands this question to evaluate how frequently
40% data manipulation in laboratory reports occurs,
allowing us to differentiate between occasional
30% never
seldom
events and habitual cheating. Although there are
20% sometimes many studies of cheating on college tests, to our
often knowledge our study is unique in examining
10% almost always
manipulation of laboratory data by
0% undergraduates.
CHM 115 CHM 335 BIO 201 BIO 100 BIO 182 BIO 193 BIO 385 Data manipulation apparently does not
Course Prefix and Number diminish as the students progress to upper
Figure 7. Results of survey, Question 9, “Do you believe this division courses or from non-major to major
course places too much emphasis on experimental results courses. Commitment to a major subject,
rather than on the processes used to get the results?” CHM
presumably because the student intends to
115, N=98; CHM 335, N=67; BIO 201, N=27; BIO 100,
N=194; BIO 182, N=40; BIO 193, N=58; BIO 385, N=66.
continue in this area of science for all or part of
N= total number of responses to the specific question. his/her professional career, apparently does not
diminish this practice.
Discussion: These results raise some important questions,
Some precautions should be taken in interpreting which include: How can this data manipulation
these findings. First, the survey was limited to be reduced or eliminated? What are the
only 7 courses at a single University, which in implications of data manipulation in the
each case were surveyed only once. We intend to undergraduate laboratory to the future careers of
survey additional courses at ASU, and hope to these students? In other words, when do the
include at least one other large university and a students stop manipulating data? In graduate,
small liberal arts college in our analysis. Second, professional or medical school? When they
the survey relies on self reporting. Some of the begin doing “real” research? When the research
students did not choose to answer all questions in is published?
the survey. The total number responding to each In response to the first of these questions, the
question in each course is presented in the figure faculty and the system itself must take significant
caption. Approximately 25% of the students responsibility. Faculty must recognize that this
chose not to answer Question 5, for example. data manipulation occurs, and not turn a blind
Third, the construction of the questions did not eye to this practice. We must examine the reason
permit us to investigate motivations other than why we require laboratory reports in the first
that the student felt his/her grade depended upon place, and whether there is another method of
the experimental results (Questions 8, 9, 15 - 17). assessing whether the student has learned the
Finally, even though students were given a clear necessary laboratory skills. Numerous laboratory
definition of “data manipulation” at the manuals are structured to provide “cook book”
beginning of the survey, it is possible that some procedures in which students are expected to
may not have clearly understood the definition of verify known biological, chemical, or physical
“data manipulation.” laws (3). However, these verification exercises
With the above caveats in mind, our results give students a false notion of the deductive
show a very strong tendency among investigative process. They begin their training
undergraduate science students to manipulate or with the preconceived notion that a “right”
make up data when writing laboratory reports. As answer exists and should be found. They are
high as these percentages are, they are similar to therefore willing to adjust their laboratory results
results observed in surveys of cheating on tests, for fear that the “wrong” answer would affect
which Cizek has described as “remarkably and their grade (4).
30
–––––––––––––––––––––––––––––––––––––––––––––– Davidson, et al., Data Manipulation in the Undergraduate Laboratory
We must change the common perception balance between process and results.
among undergraduate students that their grade The second concern, whether undergraduates
often depends upon producing the “right” answer continue data manipulation as professional
(Figure 6). This change will involve not only the scientists, has even greater implications. In the
laboratory experimental design, but also the frequently-cited study by Swazey et al., 72% of
training of graduate teaching assistants and graduate students and 59% of faculty reported to
elimination of grading based on achieving a have observed or had direct evidence of some
preconceived result. Although students must still form of scientific misconduct (6). Data
be evaluated on whether they are using falsification, however, was reported by a much
laboratory instruments accurately, we must smaller proportion of respondents, ranging from
consider whether a given laboratory can be 2% to 20%. Apparently, then, data manipulation
designed for training in the hypothetical- does decrease when the student becomes a
deductive process in addition to the specific “professional” and becomes dedicated to the
laboratory technique (4, 5). science.
Unfortunately, the number of students Over the last 5 years approximately 400
enrolled in science laboratory courses at large undergraduates at ASU have been engaged in
universities in many ways promotes cook-book research programs funded by the Howard Hughes
laboratory exercises. The limited time allowed Medical Foundation, the National Institutes of
for experiments, inability to repeat an Health and the National Science Foundation, in
experiment, and disinterest of many teaching addition to individual faculty grants.
assistants in spending adequate time to grade Conversations with these students reveal that
reports all contribute to the perception on the part once the research becomes “their own” and
of students that making up data is more profitable important to them personally, they have far less
than writing up what really happened. motivation to manipulate data, particularly if they
Faculty must rethink the reasons for have a respectful relationship with the faculty
requiring laboratory reports. If the reasons mentor. Hands-on undergraduate research
include determining whether the student was experience may therefore be important in
present, completed the tasks, understood the molding the ethical practices of students who
experiment, and learned the techniques, then the will go on to become professional scientists.
results presented here suggest that these goals are When we emphasize the importance of
not being accomplished by the current getting the “right” answer, we are teaching
mechanism of laboratory reports graded based undergraduates that their hypothesis must be
upon achieving the “right” answer. Other supported. In truth, the function of an
mechanisms for discovering whether students experiment should be to allow for a fair test of
have learned the important aspects of the exercise the hypothesis. We recognize that there exists
may include laboratory-based questions on temptation for graduate students and professional
exams, and building later experiments upon scientists to manipulate data in order to finish
earlier laboratory exercises. Instructors must be research before a deadline, to obtain the next
willing to address this problem bluntly with the grant, or to have an outstanding publication
students and teaching assistants. record. We must take serious responsibility that
At ASU some laboratories have been we do not teach data manipulation techniques at
redesigned to emphasize the inquiry approach to the undergraduate level that will continue to be
laboratories in Biology and Chemistry. Students used in later professional careers.
generate alternative hypotheses and design
experiments themselves, and concepts are Acknowledgements
introduced after, not before, results are obtained. We are grateful to Dr. Anton Lawson for helpful
Teaching assistants are trained to lead students suggestions.
into open-ended and thought-provoking
questions (4, 5). In spite of these efforts, our data Bibliography
suggest that data manipulation occurs in these 1. Cizek, GJ. Cheating on tests: How to do it, detect it, and
laboratories as well. As students commonly prevent it. L Erlbaum Assoc, Mahwah, NJ. 1999.
state, “everybody does it.” The students 2. Meade J. Cheating: Is academic dishonesty par for the
themselves overwhelmingly prefer laboratory course? Prism 1992;1: 30-32.
exercises which emphasize processes or a 3. Germann PJ, Haskins S, Auls S. Analysis of nine high

31
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
school biology laboratory manuals: promoting scientific
inquiry. J Res Science Teaching 1996; 33: 475-499.
4. Lawson AE, Lewis CM Jr., Birk JP. Why do students
“cook” data? J College Science Teaching 2000; 29(3):
191-198.
5. Lawson AE, Rissing SW, Faeth SH. An inquiry approach
to nonmajors biology. J. College Science Teaching
1990;19(6): 340-346.
6. Swazey JP, Anderson MS, Lewis KS. Ethical problems in
academic research. Amer Scientist 1993; 81: 342-553.

32
–––––––––––––––––––––––––––––––––––––––––––––– Davidson, et al., Data Manipulation in the Undergraduate Laboratory

Data Manipulation Survey

Instructions to students:
Space shuttles blow up, bridges fall, and planes crash and not all are due to natural disasters. An
undergraduate student at ASU has been conducting a research project for the last year and a half.
During his undergraduate career, he found that in some laboratory settings, there appears to be a great
deal of pressure to get the “right” result rather than an emphasis on the scientific and experimental
process. In one of his labs he found that 80% of the students manipulated data in some way during
the semester. He became concerned: where do students learn scientific ethics? Should we have faith
that human morality will overcome pressures to manipulate data in the hopes of a better grade in our
college career, or a publication in our professional career?

The purpose of this survey is to collect data on the extent to which undergraduates feel pressured to
manipulate, change, or make up data acquired in the laboratory. For example, if you only have a 30%
yield of a particular reaction, have you ever felt pressured to say you had more to get a better grade?
Moreover, how did you respond to that pressure? Alternatively, has the lab concentrated on
experimental process rather than actual results?

Data Manipulation: To change or omit acquired data or to make up data without confession to
those evaluating your performance.

1. What is your TA’s name?

2. What is you major and what year are you (freshman, sophomore, etc.)?

3. Are you:
A. Female B. Male

4. How many science labs have you taken?

A. 1 B. 2-5 C. 6 or more

5. Have you ever manipulated data or made up data in this course?

A. Almost Always B. Often C. Sometimes D. Seldom E. Never
6. If you have ever manipulated data or made up data, was it motivated by the thought of a better
grade?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

7. Have you ever observed anyone else manipulate or make up data in this course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

8. How often have you felt as though your grade in this course depended heavily on your
experimental results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

9. Do you believe this course places too much emphasis on experimental results rather than on the
processes used to get the results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

10. Have you ever manipulated or made up data in any other science course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

33
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
11. If you have manipulated or made up data, was it motivated by the thought of a better grade?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

12. If you have manipulated or made up data, was (were) the course(s):
A. Lower Division (100-200 level) B. Upper Division (300 or 400 level) C. Both A & B

13. If you have manipulated or made up data, what department was (were) the course(s) in? (Please
circle all that apply.)
A. Biology B. Physics C. Chemistry D. Zoology E. Botany F. Microbiology

14. Have you ever observed anyone manipulate or make up data in any science course?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

15. How often have you felt that your grade in a science course depended heavily on you
experimental results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

16. Do you believe that science courses place too much emphasis on experimental results rather than
on the processes used to get those results?
A. Almost Always B. Often C. Sometimes D. Seldom E. Never

17. Would you like to see lab experiments:

A. Place more emphasis on results. B. Place more emphasis on processes.
C. Have a balanced combination of both.

34
Preliminary Observations on Faculty and Graduate Student Perceptions of
Questionable Research Conduct
Ravisha Mathur, Department of Psychological Sciences, Purdue University, USA

Stuart I. Offenbach, Department of Psychological Sciences, Purdue University, USA

Keywords: Ethical dilemmas, Mentoring, Perceived misconduct, Research ethics training

When thinking about how graduate students learn the values and standards of science, most
universities and departments utilize an apprentice model. In this model, students learn values and
ethics by observing their mentor and through working with the mentor–learning via a kind of
“osmosis” process. However, the mentoring relationship between faculty advisor and graduate
student is one of the most difficult and complex relationships in academia. This sometimes
professional, sometimes personal relationship is generally beneficial to both individuals. Advisors
usually help students develop their careers and develop professionally, as well as help students
network and give them guidance with advice, support, and knowledge. Graduate students help their
advisors by assisting with projects, increasing research productivity, increasing professional visibility
through the student’s research, and can provide their mentors with personal satisfaction and a sense of
competence (1, 2). Despite this mutually beneficial relationship, vital for a graduate student’s career
in graduate school and beyond, faculty members receive very little, if any, training about mentoring.
In fact, given this lack of formal preparation, some suggest the mentoring relationship can cause as
much potential harm as it does benefits (1).
As a mechanism to transmit ethical codes and standards, the mentoring-apprentice model is,
according to some investigators, not very effective (e.g., 3, 4). In order to provide faculty and
graduate students with more effective methods of training and educating students about the
responsible conduct of research, it would be useful to determine which aspects of the practice of
research are most vulnerable to be misperceived, skewed, or violated. In this study, our definition of
the responsible conduct of research includes (but is not limited to) honesty, reporting all collected
data, using appropriate statistical analyses, and fairly recruiting research participants. Although there
is some research describing the types and frequency of scientific misconduct by faculty members and
by graduate students, there is little research examining both faculty and graduate student perceptions
of violations of the responsible conduct of research. Nor do we know how concordant or discordant
these “pairs” are. One purpose of this study was to assess these faculty and student perceptions. A
second purpose of this study was to examine the training that students receive from their faculty
advisors and departments. We hope to pinpoint how training can be improved and enhanced by
examining faculty members’ and students’ perceptions of training and regulations (at both the
department and university level).

Corresponding author: Ravisha Mathur, Department of Psychological Sciences, 1364 Psychological Sciences Building,
Purdue University, West Lafayette, IN 47907-1364, 765-494-6928 (voice), 765-496-1264 (fax),
[email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
In order to investigate these issues, we sent a
survey to faculty members and to graduate Which of the following are ways that graduate
students in each of 30 Purdue University students learn about professional values and
departments from the schools of Agriculture, ethical standards? (Circle all that apply).
Consumer and Family Sciences, Engineering, 1. Brown bag/colloquium
Liberal Arts, Science, and Veterinary Medicine. 2. Special courses devoted to this topic
Faculty members were certified to chair students’ 3. Interaction with faculty in research work
doctoral committees and graduate students were 4. Codes of ethics and professional standards
certified by the graduate school as doctoral provided by professional organizations
students. 733 faculty and 242 graduate students 5. Informal discussion of ethical problems when
received copies of the survey, and we received a they occur
total of 241 surveys from faculty (of which 225 6. Department policies for teaching and research
contained usable data) and 47 surveys from 7. Discussion of ethics and values in regular
students (all of which were usable data).1 course work
Although the participation rate in this survey was
comparable to previous research on similar issues Figure 1: Item 2 from Part 1 of the Survey
with mail-in surveys (e.g., 5), we were
disappointed that we did not receive more addressed how information about the responsible
responses from students (which limited the conduct of research is exchanged (Item 2 of Part
analyses and results reported below). The 1 is shown in Figure 1). The questions in Part 1
distribution of returns by Gender and by focused on how and where students learned about
Discipline are in Tables 1 and 2, respectively. the responsible conduct of research and if
students and faculty knew of others who had
been involved in ethical conflicts. The main
Female Male section of the survey, Part 2, consisted of 38
Faculty 47 162 hypothetical dilemmas (each included a proposed
Grad. Student 16 29 action to resolve the dilemma). The dilemmas
Table 1: Number of responses by gender were written to cover the following types of
problems (which were supported the
The percentage of responses from both male and confirmatory factor analysis described below):
female faculty members and graduate students 1) Information Sharing in the Lab;
matched the gender distribution for the entire 2) Truth/ Completeness in Writing up Research
faculty (faculty: 22% female and 78% male; Results;
graduate student: 35.5% female and 64.5% 3) Misleading the Audience (Plagiarism);
male). Equivalent comparisons of responses 4) Seeking Credit for doing the Research; and
from the different disciplines were more difficult 5) Consent Issues.
to make since different numbers of departments (Examples of the dilemmas for each factor are
from each discipline were asked to participate. shown in Figure 2.) Participants responded by
As Table 2 indicates, more responses were rating each dilemma on a five point Likert scale
received from the Schools of Agriculture, (Strongly Disagree to Strongly Agree). The third
Engineering, and Science. Only a few graduate and final section of the survey examined
students from Consumer and Family Sciences participant’s perceptions of university and
and from Liberal Arts participated. Most of the departmental policies on the responsible conduct
student responses were from Agriculture and of research and whether the faculty member or
from Engineering. graduate students would feel comfortable
There were three parts of the survey. Part 1 reporting incidents of suspected misconduct.

Agriculture CFS Engineering Liberal Arts Pharmacy & Science

Medical Sci.
Faculty 52 23 32 27 20 38
Grad. Stud. 13 4 10 2 7 7

Table 2: Number of Responses by School

36
–––––––––––––––––––––– Mathur & Offenbach, Faculty and Graduate Student Perceptions of Questionable Research Conduct

a. Sharing Information: Grant is in his office one day and sees his officemate’s lab notebook
open. While paging through the notebook, he discovers that Anli has found a way to metabolize ABC
enzyme. Grant has been working for two months to discover a way to metabolize this enzyme for his
dissertation. After thinking about it for a few days, Grant decides to use the same process to keep his
dissertation on track. He does not bother to tell Anli because she is in his lab group and probably would
not mind anyway. Do you agree with his decision?
b. Writing: Mei has been collecting data for a long-term study for the past two years. Although she
still is in the middle of the data collection phase, the trends she sees in her data are very exciting. She
decides to write up her results and present them as a complete study and continue to collect data for the
full term of the study. She plans to publish those data in at least two “follow-up” reports. Do you agree
with her decision?
c. Misconduct: Angelo has written an article in which he included a full paragraph from a paper
written by a student for a class Angelo was teaching. Do you agree with Angelo’s decision to include the
paragraph?
d. Seeking Credit: John has written an article in which he included a full paragraph from a
pre-publication version of an article reviewing the research literature in his area of interest. The author of
the article was planning to submit it to a prominent journal that publishes such reviews. Do you agree
with John’s decision to include the paragraph?
e. Consent Issues: Professor Gleeson is conducting a research project concerned with social
customs in a village in rural South Africa. The village consists of members of a single tribe, and is led by
a tribal chief and council of elders who make all decisions for the village. The tribal chief insists that he
will decide if his villagers can participate in Professor Gleeson’s research project, and that he (the Chief)
will distribute the payment to the villagers. Professor Gleeson may not ask the villagers whether they
want to participate because that would be an insult and challenge to the Chief and Elders of the village.
Do you agree that Professor Gleeson can go ahead with the research project if the Chief and Elders
approve?

Figure 2: Sample Hypothetical Dilemmas from Part 2 of the Survey

(Two of these items are shown in Figure 3.) believed supportive faculty members provided
Items from both Part 1 and Part 3 were adapted such information. Sixty-seven percent of faculty
from Judith Swayze and coworkers’ survey of members believed professional organizations
faculty and students (6). Items for Part 2 were provided such information compared to only
written by the authors and were based on real 15% of graduate students (t = 28.377; Only t-
events and scenarios gleaned from reading and values significant at .05 or less are reported).
teaching about the responsible conduct of This difference probably reflected a lack of
research for the past five years. contact with such organizations by graduate
Participants were given a response sheet to students. Graduate students also relied more on
use as their answer sheet and were asked to other students as a source of information (51%),
return the response sheet in a self addressed a source not considered by faculty members
envelope we provided them. Once we received (15%, t = 16.97).
the survey, a third party removed any identifying Interactions with faculty in research work
information. The responses on each survey form and informal discussions of ethical problems
were entered into a computer file separately by were considered effective settings to learn
the two authors. All coding errors then were professional values by 90% or more of students
reconciled by the authors. and faculty. Brown bag discussions, colloquia,
and courses, on the other hand, were not seen as
Results effective settings by most respondents
(percentages all less than 30%).
Part One. The first questions focused on
We also asked whether respondents ever
settings in which respondents learned some or all
discussed with peers value issues related to
of their professional values. Seventy-two percent
external sources of research funding or the
of faculty members and 60% of graduate students
application of research findings. Eighty percent
37
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

2. How active a role does your department actually take in preparing graduate students to recognize
and deal with ethical issues in your field as part of their training?

Very Active Somewhat Not very Not at all

active active active active active

3. Could you report cases of suspected misconduct in your department without expecting
retaliation?

Misconduct by a faculty member: Yes definitely, Yes, Probably, Probably not, Definitely not
Misconduct by a graduate student: Yes definitely, Yes, Probably, Probably not, Definitely not

Figure 3: Items #2 and #3 from Part 3 of the Survey

of faculty members and 47% of the graduate Although we could not adequately examine
students (t = 18.263) did so. In addition 38% of the faculty-student differences on the responses
faculty members and 11% of graduate students to the Hypothetical Dilemmas because of the
actually knew someone who had refused to disparity in the number of responses from each
participate in a research project because of group, we were able to draw some tentative
personal reservations about funding sources. conclusions. Faculty members clearly took
These faculty-student difference probably “more extreme” views than did students. That is,
reflects differences in age and experience in the faculty members were more likely to indicate
field. strong disagreement or agreement with the action
What is clear from these analyses is that taken in a dilemma than were graduate students.
faculty members and students do have different For example, on the 20 dilemmas that
views of the best place or way to learn about contributed to the five factors, more faculty
professional standards and to learn to recognize members responded “strongly agree” (or
ethical research issues. “strongly disagree”) on every dilemma.
Part 2: Hypothetical Dilemmas. A Graduate students had more moderate responses.
confirmatory factor analysis of the hypothetical Actually, there were no faculty-student
dilemmas produced five factors: 1) Information differences in the number of combined “strongly
Sharing in the Lab; 2) Truth/ Completeness in agree” and “agree” (or “strongly disagree” and
Writing up Research Results; 3) Misleading the “disagree”). Thus for the second item in Figure
Audience (Plagiarism); 4) Seeking Credit for 2, of the 98% faculty members who disagreed
doing the Research; and 5) Consent Issues. The with the action, 80% checked “strongly
alphas for these variables were moderate, ranging disagree.” All of the graduate students disagreed
from .47 - .61. We recognize that not all of the with the action, but only 43% expressed strong
dilemmas applied equally to all of the disciplines disagreement. Perhaps faculty members’ greater
sampled in this survey, but we were pleased that experience with ethical issues has led them to be
some general factors appeared. The nature of more certain of their views (or the students’ lack
the five factors can be explained in several ways. of experience led them to be more tentative).
First (and probably foremost) is the construction Finally, while the responses to the
of the scenarios by the principle investigators. hypothetical dilemmas made intuitive sense, the
Construction of these scenarios was not a random construction of the dilemmas is more complex
process, and the factors extracted from the than we thought. Respondents often commented
analysis may simply confirm biases and that they saw some items as dealing with
predispositions that entered into our construction multiple ethical issues or that there was not
of the items. On the other hand, the areas enough information presented to make a
represented by the five factors have been judgement. This may be one reason alpha levels
identified by many investigators as areas of were low for the five factors. More thought must
concern vis-a-vis research ethics. The fact that go into the development of items that have a
these items hang together at all may be a more specific focus (and are less complex) for a
confirmation of the concerns many investigators survey of this type.
and ethicists have about the process of research. Two sets of analyses were not computed.
38
–––––––––––––––––––––– Mathur & Offenbach, Faculty and Graduate Student Perceptions of Questionable Research Conduct
Analyses to compare factor scores for students misconduct by a faculty member or by a
with those of faculty were not conducted because graduate student without expecting retaliation.
the factor scores have not yet been corrected for The results in Table 6 show that 89% of faculty
directionality differences. That is, some factors members believed they could report misconduct
include items with which most respondents agree by a graduate student “safely.” They would
and items with which most respondents disagree. expect no retaliation. The graduate students also
The point values for these items needs to be on seemed less concerned about retaliation if they
the same scale or have the same valence in order reported misconduct by another student.
to examine factor scores. The other analyses not Seventy-three percent thought it was safe to
yet conducted would have compared student report misconduct by another graduate student.
responses with those of their mentors. These Reporting misconduct by faculty members was
analyses depended on both the student and his or another matter. Fewer faculty members were
her mentor actually submitting a survey, and comfortable about reporting misconduct by a
having the student identify his or her mentor. colleague (73%). Only 55% of students thought
Unfortunately, we were able to identify only five they could report misconduct by a faculty
faculty-student pairs, precluding any analysis of member “safely.” In contrast, 28% of the faculty
whether the two are concordant or discordant. members who responded said they would not
feel safe reporting misconduct by a faculty
Questions about department and colleague. Almost half of the graduate students,
university policies 44%, were concerned about retaliation for
The questions in Part 3 focused on respondents reporting a faculty member’s misconduct. These
perceptions of the role that departments should results seem consistent with anecdotal data. A
take and actually do take in preparing students to cursory review of comments from the electronic
recognize and deal with ethical issues (see Tables list-serve Sci-Fraud reveals a concern by many
3 and 4). Significantly more students than participants that to make a good faith allegation
faculty (70% vs. 45%) reported almost no effort that a faculty member has engaged in
by their departments to train them to recognize misconduct is to place one’s career in jeopardy.
and deal with ethical issues in science (it also is Finally, we asked about knowledge of
interesting that 16% of faculty members thought university and departmental policies on
their departments were active, but only 6% of the misconduct. Half of graduate student
students shared that perspective). Thus both respondents did not know that the University has
faculty and students believe academic a research misconduct policy and 72% do not
departments should take a more significant role know if their department has such a policy. The
in training graduate students to recognize and faculty were more knowledgeable – 63% knew
deal with ethical issues (we only asked about there was a university policy. However, only
academic departments, faculty members and half of them were familiar with the policy’s
students may actually ascribe greater contents.
responsibility to larger academic units — e.g.,
schools, graduate school, etc.).
There is a mismatch here –
faculty and students wanting Some-
Very what Not very Not at all
departments to take a role and Active
active active active
departments not doing that. And active
there is no formal structure at Faculty 37 45 14 03 01
the university level for training Grad. Stud. 22 52 22 04 00
in the responsible conduct of
Table 3: Role a department should take (percent agreeing)
research. Thus, the student is
left to his or her own devices. Some-
The most frequent choice made by Very Not very Not at all
Active what
active active active
students seems to be to ask active
another student or to ask the Faculty 02 14 38 34 11
advisor. Grad. Stud. 02 04 26 51 17
The next two questions asked
whether one could report Table 4: Role a department does take (percent agreeing)
39
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Misconduct Definitely Probably Probably Definitely

Respondents
by Yes Yes Not Not
Faculty
Members
} Faculty
Students
32
48
41
41
23
09
05
01
Graduate Faculty 04 51 40 04
Students } Students 11 62 23 04

Table 5: Reporting Misconduct Responses (percentage agreeing)

Conclusions:
The hypothesis that graduate students learn to
Notes
1. These numbers differ from the totals in Tables 1 and 2
identify and deal with ethical situations in as some participants did not answer the gender or
research from their mentors without specific discipline questions.
instruction or discussion could not be tested
using the data collected in the first “pass” of our Bibliography
study. We received too few mentor-student data 1. Johnson, W. B., & Nelson, N. (1999). Mentor-protégée
pairs to make any analysis. Our failure to obtain relationships in graduate training: Some ethical
data relating mentor’s values directly to that of concerns. Ethics Behav, 9, 189-210.
their specific students was disappointing – only 2. Wright, C. A., & Wright, S. D. (1987). The role of
five student - mentor pairs were identified (we mentors in the career development in young
hope to rectify this situation by increasing the professionals. Fam Relations, 36, 204-208.
size of the student data pool). However, we 3. Brown, S., & Kalichman, M. W. (1998). Effects of
training in the responsible conduct of research: A survey
believe the modeling or osmosis hypothesis
of graduate students in experimental sciences. Sci Eng
probably will not be supported because of the Ethics, 4, 487-498.
different perceptions graduate students and 4. Folse, K.A. (1991). Ethics and the profession: Graduate
faculty members have of how scientific values school training. Teach Sociol, 19, 344- 350.
are transmitted. Faculty members and students 5. Mearns, J., & Allen, G. J. (1991). Graduate students’
do rely on other faculty members, but only the experiences in dealing with impaired peers, compared
students rely on their student peers. At the same with faculty predictions: An exploratory study. Ethics
time, both faculty and students believed that Behav, 1, 191-202.
interactions in the work or lab settings would be 6. Swazey, J. P., Anderson, M. S., & Louis, K. S. (1993).
Ethical problems in academic research: A survey of
useful in learning to recognize and deal with
doctoral candidates and faculty raises important
ethical situations. Unfortunately, this expectation questions about the ethical environment of graduate
means that people seem to want to learn from education and research. Am Scientist, 81, 542-553.
“personal experience,” but no one wants to have
that kind of personal experience.
One thing is certain, things will not continue
in the same way. Actions by the National
Institutes of Health to require specific education
on the responsible conduct of research generally
specifically will require universities to do a better
job. That better job might be facilitated with a
more complete understanding of how students
are learning now and by determining not only
what they are learning , but also by determining
what they are NOT learning.

40
Constructing a Personal Model of Research: Academic Culture and the
Development of Professional Identity in the Professorate
L. Earle Reybold, Adult and Higher Education, The University of Texas at San Antonio, USA

Keywords: Academic culture, Faculty development, Graduate socialization, Professional identity

, Research
ethics, Research training, RCR

Doctoral students pursuing academic careers are educated in awkward and mostly tacit
apprenticeships. As students, they are expected to learn professional knowledge and the technical
skills associated with their program of study. Yet, they must simultaneously absorb the culture of
academe and learn their future roles as faculty members. Because learning and thinking are situated
in a social milieu, socialization is a process initiated and established in contexts that construct
knowledge through activity (1). In other words, academic culture and educational knowledge “act
together to determine the way practitioners see the world” (p. 33).
Generally, socialization studies have investigated academic culture as context for student learning
and development. Many of these studies focus on the social aspects of academic culture, particularly
relationships between students and their colleagues or professors (2, 3, 4, 5). These socialization
studies concentrate on students’ experiences as students in higher education and are centered on
classroom modality.
Likewise, inquiry into new faculty socialization segregates faculty roles and responsibilities into
particular genres of experiences such as teaching success (6) and tenure and promotion processes (7).
Unfortunately, faculty socialization studies fail to address how graduate school experiences,
particularly as they are situated in an academic culture, affect the development of professional
identity and ultimately professional decision-making and activity.
When the concept of professional identity and competency is addressed in the faculty
socialization literature, the discussion surveys the development of the faculty teaching roles but
ignores the complex faculty identity as teacher, researcher, and service provider. This lack of
attention to an integrated identity that begins to emerge during graduate studies portrays faculty
socialization in perfunctory terms. For example, Boice discusses new faculty success in terms of
teaching style and mastery (6). The author notes the characteristics of “quick starters,” but these are
teaching characteristics of new faculty, with no attention to the development of these characteristics.
Pollard, Pollard, & Rojewski also investigate the college teaching experience of new faculty (8).
They argue that doctoral students are academically prepared for their careers in higher education, but
their study concentrates only on the impact of higher education culture on new faculty.

Purpose of Study and Research Focus

The purpose of this study is to describe the role of academic culture in determining a personal model

Corresponding author: L. Earle Reybold, Adult and Higher Education, College of Education and Human Development, The
University of Texas at San Antonio, 6900 North Loop 1604 West, San Antonio, TX 78249-0654, 210-458-5429 (voice), 210-
458-5848 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
of ethical research practice in the professorate. members at three universities in two states, and
While little is known about the construction of through archival data such as program materials
faculty identity and role expectations during and reflection journals supplement the interview
graduate studies, even less is understood about data. Interviews were conducted using a semi-
the impact of student experiences on professorial structured format to allow comparison of data
activities and decision-making, particularly across participants (11). In general, interview
research competence and reasoning. Two questions addressed student and professional
questions demand consideration. First, how are identity, academic culture, training in teaching
doctoral students socialized into the practice of and research, and ethical decision-making as a
academic research? Further, how do these professional. Journaling allowed students to
students construct a model of research standards explore and document their process of decision
and ethics that will inform their future practice as making as relevant issues arose, the entries were
faculty members? guided by the following statement: Describe your
Two general assumptions guide this inquiry: decisions that are most important to your
• Socialization into the professorate is a preparation for the professorate.
developmental rite of passage rather than
two discrete phases of socialization marked Standards for Quality Research
by graduation and/or faculty appointment.. Emphasizing qualitative inquiry as a creative
• Preparation for the professorate is situated in process, Patton (10) reminds researchers of the
“technical side to analysis that is analytically
an academic culture that shapes one’s
rigorous, mentally replicable, and explicitly
personal understanding of the professorate systematic” (p. 462). Merriam (9) adds that
and professional identity and perceived roles. qualitative research findings “are trustworthy to
This study initiates a two-phase longitudinal the extent that there has been some accounting”
qualitative investigation. Using case study (p. 198) for quality. In general, the criteria for
methods (9), this study focuses on doctoral trustworthy qualitative research include rigorous
students’ perceptions of research ethics in and systematic data collection and analysis
education. Interview questions concentrated on techniques, credibility of the researcher, and
emergent definitions of research ethics, training belief in naturalistic inquiry (10). The quality of
in research ethics, and experiences of ethical this study is enhanced by several factors. First, I
dilemmas. have experience as a qualitative researcher and
Case study research is uniquely geared have taught qualitative methods at the graduate
toward description and understanding of level. Further, triangulation of methods and peer
institutional culture and its impact on review of data and analysis will enhance the
perspective. Merriam describes case study trustworthiness of the data. Finally, the multi-site
research as an ideal design for exploring design encourages usability of the findings
participants’ understanding and perspective (9). beyond the university settings included in the
Further, she says case study is appropriate when study.
inquiry is interested in “process rather than
outcomes, in context rather than a specific Situating Faculty Identity Development in
variable, in discovery rather than confirmation” Academic Culture
(p. 19).
This study is framed by the concepts of research
Sampling for this phase of the study is
ethics and integrity, faculty socialization and
network sampling, which locates participants
enculturation, and professional identity
through recommendations of initial participants
development.
and key informants based on selected criteria
(10). Participants were located at three
Research Ethics and Integrity.
universities in Georgia and Texas, including
institutions identified as Research I, Research II, Research is often messy and complicated. Best-
and Doctoral II. Participants were doctoral case scenarios of theoretical contributions and
students in education preparing for a faculty improvement of practice are weighed against
career in academe. questionable issues of right and wrong research
Data were collected through in-depth behavior. In these cases, research decisions may
interviews with doctoral students and faculty evolve as uneasy guesses with no obvious
consequence. Confronted with uncertain choices,
42
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Reybold, Constructing a Personal Model of Research
how do researchers define and respond to ethical individual’s values and beliefs to fit the “cultural
dilemmas? ethos of the institution” (p. 25). Tierney and
Ultimately, ethical decision-making reaches Bensimon continue this emphasis on
beyond the local boundaries of specific research socialization in academe, focusing on the tenure
projects. Because research is fundamental to process as the locus of organizational
higher education, it could be argued that research socialization (7). Although they offer strategies
decisions symbolize the moral character of for anticipatory and organizational socialization,
higher education. Under the guise of exploration the authors do not focus their attention on the
and discovery, research is a noble enterprise. But transition process.
research agendas are realized within the Bergquist examines academe within the
“publish-or-perish” mentality of higher education framework of organizational culture, concluding
in which ethical dilemmas may become that there are four distinct cultures: collegial,
stumbling blocks to promotion and tenure. This managerial, developmental, and negotiating (13).
is the context where doctoral students are Culture, he says, “provides meaning and context
socialized toward the professorate; this is the for a specific group of people,” adding “the
culture that trains future faculty members as culture holds the people together and instills in
future researchers. them an individual and collective sense of
purpose and continuity” (p. 2). Further,
Faculty Socialization and Enculturation. Bergquist says culture defines the nature of
Tierney & Rhoads (12) remind us that reality for members of a given culture, providing
“organizations exist as social constructions” (p. the “lenses through which its members interpret
1) that revolve around shared understandings. and assign value to the various events and
This organizational culture shapes behavior and products of the world” (p. 2). Although there are
expectations, bounding faculty socialization. four distinct cultures within academe, one will
Tierney & Rhoads define faculty socialization as usually be dominant. Bergquist notes that the
“the process through which individuals acquire interaction among the four unequal cultures helps
the values, attitudes, norms, knowledge, and “to produce the often confusing and paradoxical
skills needed to exist in a given society” (p. 6). conditions in which contemporary faculty find
Their definition of faculty socialization as themselves” (p. 7).
transmission of culture complements this study Both Bergquist (13) and Tierney & Rhoads
of professional identity development. (12) note the influence of academic culture on
Tierney & Rhoads (12) describe academic faculty perspectives, decisions, and behavior;
culture as the nexus of five forces: national, also, they agree that cultural differences create a
professional, disciplinary, individual, and backdrop of conflict for members within a given
institutional. Although these are conceptualized culture. This study extends their conclusions to
as distinct subcultures, these forces are graduate education, adding that students also are
synergistic and do not operate independently of influenced by academic culture. Further, the
one another. Professional identity is an aggregate transition process from doctoral studies to the
sense of self that develops across these professorate adds another layer of possible
subcultures. This process of socialization occurs conflict between academic cultures.
in two overlapping stages: anticipatory
socialization and organizational socialization. Developing a Professional Identity.
The anticipatory stage “pertains to how non- Marcia defines identity development as a self-
members take on the attitudes, actions, and constructed organization of drives, abilities,
values of the group to which they aspire” (p.23). beliefs and individual history (14). Bruss &
The organizational stage, on the other hand, Kopala (15), building on Marcia’s definition,
involves initial entry and role continuance. define “professional identity “the formation of an
Noting the importance of the transition process, attitude of personal responsibility regarding one’s
Tierney & Rhoads comment that when role in the profession, a commitment to behave
anticipatory socialization and organizational ethically and morally, and the development of
socialization are consistent, the socialization feelings of pride for the profession” (p. 686).
process is affirming. When socialization This definition directly connects professional
experiences are not consistent, the organization identity to professional behavior.
will attempt to modify or transform the While the identity development literature is
43
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
concerned predominantly with the psychological professorate as a delicate balance of professorial
aspects of self, identity may be viewed as both roles, most participants emphasized the
personal and social. Social identities result in preeminence of becoming a researcher, and only
identity relationships within a given culture, and one participant noted a teaching role being more
these identity relationships determine identity important than a research role. For example,
status and role expectations (16). For the Betsy says, “research is painful and boring, but
purpose of this study, status and role expectations the doctorate is about what the university
will be examined as cultural aspects of considers important—getting published!”
professional identity development, particularly as Echoing this sentiment, Claire says the “doctoral
they relate to anticipatory socialization during degree is mainly trying to get us into the research
the graduate school experience (7). part of being a professor and much less teaching;
Graduate training is expected to nurture the it is indoctrination into the research aspect of
development of professional identity. In their being a professor.”
discussion of psychology students, Bruss and While some participants came in with
Kopala (15) described graduate school training as considerable research experience, most are
professional infancy and “the training institution . concerned that they don’t “know what to do with
. . as an environment wherein it is the task of the the research” after the dissertation process. Post-
faculty and training staff to nurture and promote dissertation concerns include translation of
growth” (p. 686). However, academic culture is theory into educational practice, establishing a
not always nurturing; structural problems in research agenda, and getting published.
graduate education are potentially harmful to What are the perceived standards for ethical
students’ self-esteem (17). Attitudes—good and research and who defines ethics in academic
bad—about professional responsibility, ethical settings? Coursework in research ethics is almost
behavior, and professional pride are constructed nonexistent. As students, participants expect
within the cultural context of graduate training. professors to guide them through the process of
These attitudes produce social identities and role learning and implementing ethical research, but
expectations that persist through a graduate they are relying instead on their own sense of
student’s transition into the professorate. In right and wrong. Julia says she relies on her
short, academic culture exerts directive force “internal gyroscope” to guide her decisions; and
over professional decision-making and activities. Claire relies on her “personal ethics and personal
Chickering & Reisser, in their study of morals.” Grace adds that “ethics is about power
college student development, define identity as a differences.” Her professors talked about
sense of self (18). The process of identity collaboration and high quality, but their practice
development results in “a solid sense of self expressed a disregard for the Institutional Review
[that] emerges, and it becomes more apparent Board (IRB), quality research, and research
that there is an I who coordinates the facets of integrity.
personality, who ‘owns’ the house of self and is More than a lack of definition of ethical
comfortable in all of its rooms” (p. 49). research, participants are concerned and confused
about “grey” areas of research ethics and believe
Findings they must define ethical research according to
To describe the role of academic culture in their own experiences and standards.
determining ethical research practice, data were Interestingly, the two participants with training in
analyzed within four concentrations: the medical ethics find research ethics easier to
perceived role of research in higher education, define. The other participants have scattered
the perceived standards for ethical research, the definitions of research ethics, with most
actual ethical dilemmas experienced by graduate positioning ethical research as a data collection
student researchers, and the factors that hinder or and/or analysis issue. However, a couple of
support ethical research. participants have a complex, comprehensive
What is the perceived role of research in definition of research ethics, including researcher
higher education? Participants in this study attitude and choices throughout the research
experience research and subsequent publication process. One participant noted that article
as an institutional priority and a personal badge readers have an ethical responsibility to read the
of prestige. While one participant views the results thoroughly. Another participant, Grace, is
quite concerned with the power issues that
44
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Reybold, Constructing a Personal Model of Research
impact ethical decision-making: “power issues presentations because funding for student travel
come into play, whether we like to admit it or to conferences depends on authorship. Grace did
not…these are times we just have to make a try to confront the professor, but to no avail. The
mental note, ‘this is not right’…. But I’m at a professor was on the editorial board of the
point where I have no power to address this.” journal that published the article, and she
One participant has a collaborative believed the issue would not be taken seriously.
relationship with her major professor. Kelly says Participants report that even when the research
her discussions with her major professor about situation is uncomfortable, they “don’t want to
research methods and ethics have been sacrifice the relationship” by removing
invaluable, even to the point where she feels themselves from the project.
comfortable mentoring other students with Another type of dilemma involves committee
research problems. Although Betsy claims to make-up. One participant had approval for a
have a collaborative and mentoring relationship mixed design dissertation, but her committee
with her major professor, she often finds herself politicized her design and held up her research.
involved in ethical dilemmas with others in the She decided “to write it one way for the
same department. For the participants in this dissertation” and then publish it using her mixed
study, the most beneficial contribution to ethics design approach. Other participants experienced
and methods training is involvement in actual negative “shaping” of their research based on
research projects, particularly pilot studies of professors’ interests. As one participant reports,
own research and collaborative efforts as “professors stay in their comfort zones” and
research partners with professors, but only when won’t head committees outside their personal
that contribution is valued and rewarded as equal. interests. This is particularly problematic in
What types of actual ethical dilemmas do small departments with few faculty members.
graduate student researchers experience? While What factors hinder or support ethical
most participants define ethical dilemmas in research? Several factors hinder ethical
terms of research methods, their experiences of research: institutional/structural, relational/
ethical dilemmas focus more on relationships and positional, and technical. First, the culture of
issues of power and coercion. One participant academe encourages ambivalence toward the
reports her professor “uses” students to review issue of ethical research. Institutions reward
his own material prior to publication. Student research productivity, even at the expense of
assignments in non-related courses revolve other professorial roles, perpetuating the adage,
around this professor’s research agenda, and publish or perish. While some professors
students are expected align their work to match “nudge” their students to commit ethical
that agenda. Several participants report being violations, others ignore the need for training and
forced to manipulate data to yield desired guidance in ethical research practice. Dan,
outcomes; if a student refuses, he or she is no looking toward a future career in academe,
longer funded as a research assistant. Kelly, a acknowledges that “tenure is political, so go way
research assistant on a grant-funded study, voiced beyond their expectations!”
disapproval of research decisions being made by A second factor hindering ethical research is
professors on the grant: the role of hierarchy in academic relationships.
I’ve been vocal, but I wasn’t a threat or Graduate students are afraid to report ethical
anything. I was unhappy with the way the violations; they fear losing their assistantships
professors were doing things . . . . I was just and professorial support. As a student, one
going along, and it hit me. Did I feel free to participant notes that “it’s important to know
leave? No! To a certain extent, this is part of where your allegiances lie; the only way you’ll
being a graduate student. I mostly feel free to
voice my concerns, but in this case, it was an
get lobbied for is if you are clearly in someone’s
ultimatum—or I was off the grant! I never want camp.” Only one student, Kelly, says her
to do this in my own research. professors treat her as a peer. Her major
Another participant, Grace, reports working on professor, she says, “got me involved with his
presentations and articles with more than one projects, but told me to ‘find your own thing—
professor and negotiating authorship—but the academia isn’t just doing other people’s work.’”
articles were published without her name or with Several participants alluded to expecting a
a different authorship order than negotiated. This similar role as junior faculty; coercion will
is particularly troublesome at conference continue to force them to make ethical decisions
45
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
that might not be supported by academic much of higher education focuses on the
expectations. development of technical competence and critical
A third factor that hinders ethical research is thinking skills, the transformation from student
the lack of training and exposure to guidelines. to faculty member is too often left to chance.
Only those participants with medical Future inquiry will explore the development
backgrounds had any courses in ethics, and those of professional identity throughout preparation
courses dealt with medical ethics rather than for the professorate, and how this emerging
research ethics. Only one participant reports identity impacts professional decision-making as
research ethics discussed in her doctoral research a scholar.
classes. None of the participants in this study
knew of any guidelines for education research Bibliography
practice other than IRB guidelines. 1. Brown JS, Collins A, Duguid P. Situated cognition and
Only one participant, Kelly, reports the culture of learning. Educational Researcher 1989;
preparation for ethical research. Her major 18(1):32-42.
professor and a network of research professors 2. Heinrich, KT. Doctoral advisement relationships
between women. Journal of Higher Education 1995;
provide guidance through formal and informal
66(4):447-469.
mentoring and involvement in various research 3. Martinez Aleman AM. Understanding and investigating
projects. This particular participant has female friendship’s educative value. Journal of Higher
published with several professors, and her Education 1997; 68(2):119-158.
contributions are valued as equal to those of the 4. Schroeder DS, Mynatt CR. Female graduate students’
professors. In fact, this professor reminds the perceptions of their interactions with male and female
participant that she is his “primary major professors. Journal of Higher Education 1993;
responsibility” and that she is to develop her own 64(5):555-573.
line of research separate from his. Another 5. Stalker J. Athene in academe: Women mentoring women
in the academy. International Journal of Lifelong
participant feels secure in her relationship with
Education 1994; 13(5):361-372.
her major professor, but says her other 6. Boice R. Quick starters: New faculty who succeed. New
experiences with faculty members in the same Directions for Teaching and Learning 1991; 48:111-121.
department make her bitter and wary. She notes 7. Tierney WG, Bensimon EM. Promotion and tenure:
there are two levels of culture in the department, Community and socialization in academe. New York:
and a “lot of politics gets played below the State University of New York Press; 1996.
surface” even though “we represent ourselves as 8. Pollard C, Pollard RR, Rojewski JW. The college
a united front.” teaching experience: A qualitative study. College
Student Journal 1991; 25(2):149-154.
9. Merriam SB. Qualitative Research and Case Study
Summary and Conclusion Applications in Education. San Francisco: Jossey Bass;
Almost all participants in this study raised 1998.
themes of power and censorship. The impact of 10. Patton MQ. Qualitative evaluation and research
coercion and fear on the research process must be methods. 2nd ed. Newbury Park, CA: Sage; 1980/1990.
explored. Graduate students believe their 11. Bogdan RC, Biklen SK. Qualitative research for
research is censored on several levels: personal, education: An introduction to theory and methods. 2nd
institutional, and systemic. First, graduate ed. Boston: Allyn and Bacon; 1992.
students expressed fear of retaliation if they 12. Tierney WG, Rhoads RA. Faculty socialization as a
resisted their faculty advisor’s management of cultural process: A mirror of institutional commitment.
ASHE-ERIC Higher Education Report No. 93-6.
their research. Further, these students believe Washington, D.C.: The George Washington University,
they are bound by the dissertation committee School of Education and Human Development; 1993.
structure and the institutional support of highly 13. Berquist WH. The four cultures of the academy. San
productive faculty members. Finally, censorship Francisco: Jossey-Bass; 1992.
is systemic, according to these students’ 14. Marcia J. Identity in adolescence. In: Adelson J, editor.
experiences, because certain topics are Handbook of adolescent psychology. New York: Wiley;
“favorites” of funding agencies. Likewise, these 1980. p. 159-187.
students believe journal editors and blind reviews 15. Bruss KV, Kopala M. Graduate school training in
control the emergence of new knowledge. psychology: Its impact upon the development of
professional identity. Psychotherapy 1993; 30 (4); 685-
The goal of higher education is the 691.
preparation and personal development of
competent, well-trained professionals. While
46
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Reybold, Constructing a Personal Model of Research
16. Goodenough WH. (1969). Rethinking ‘status’ and
‘role’: Toward a general model of the cultural
organization of social relationships. In: Tyler SA,
editor. Cognitive anthropology. New York: Holt,
Rinehart and Winston; 1969. p. 311-30.
17. Egan JM. (1989). Graduate school and the self: A
theoretical view of some negative effects of
professional socialization. Teaching Sociology 1989;
17, 200-208.
18. Chickering AW, Reisser L. Education and identity. 2nd
ed. San Francisco: Jossey-Bass; 1993.

47
Undergraduate Academic Cheating as a Risk Factor for Future Professional
Misconduct

Julio F. Turrens, Department of Biomedical Sciences, University of South Alabama, Mobile,

AL, USA

Irene M. Staik, Department of Behavioral and Social Sciences, University of Montevallo,

Montevallo, AL, USA

D. Kristen Gilbert, Department of Behavioral and Social Sciences, University of Montevallo,

Montevallo, AL, USA

W. Curtis Small, Department of Biomedical Sciences, University of South Alabama, Mobile,

AL, USA

John W. Burling, Department of Behavioral and Social Sciences, University of Montevallo,

Montevallo, AL, USA

Keywords: Academic misconduct, Cheating, Undergraduate education

“T’is education forms the common mind:

Just as the twig is bent the tree’s inclined.”
Alexander Pope (1688-1744)

Scientific misconduct may be more acceptable in the minds of those professionals who grew
accustomed to lower academic standards during their formative undergraduate years. The hypothesis
proposed in this paper is that the recent increase in cheating at the undergraduate level is likely to
result in an increase in the number of future professionals involved in scientific misconduct.
Twenty years ago, academic misconduct at the undergraduate level was considered by the great
majority of both students and faculty as unacceptable and dishonest behavior. Currently, not only are
most undergraduate students aware that misconduct is very common but most of them by their Junior
year have participated or witnessed more than one event. Even those students who do not engage in
academic misconduct have become more skeptical of the need to be personally responsible for their
own academic work and accept this lowering of standards as a fact of life.
Because of these changes in the environment of higher education, the incidence and prevalence of
cheating by college students has been an area of intense concern for educators and researchers since
the 1970s. A vast number of articles in the literature indicate that cheating or academic dishonesty is
at epidemic proportions within academia (1-7). A representative sampling of articles documenting this
Corresponding author: Julio F. Turrens , Department of Biomedical Sciences, University of South Alabama, Mobile, AL
36688, 251-380-2714 (voice), 251-380-2711 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Eric # or Journal Year Sample size Institutions Reported cheating

ED427355 1998 203 four years 78 %
two years 57%
EJ351071 1986 380 > 50%
ED334921 1990 232 Rutgers 88%
ED347931 1992 87 81%
EJ449186 1992 6000 31 top-ranked business: 87%
engineering: 74%
science: 67%
humanities: 63%
EJ489082 1994 480 2 colleges 89%
EJ518822 1995 300 83%

Res. High Ed. 1984 380 mid size 54.1%

37:487-502, 1996 1994 474 liberal arts 61.2%
Table 1. Studies showing increased cheating by undergraduate students.

recent increase in cheating by students is shown potential disagreement and/or confusion as to

in Table 1. Estimates in the literature reveal that what constitutes cheating behaviors in and out of
75% to 98% of college students cheat at least the classroom, the students and faculty of the
once during their college career (8, 9). Students, University of Montevallo were presented with a
also reported that they are deterred from cheating variety of examples of academic misconduct, and
only by a fear of getting caught and public then asked to rank their perceived severity on a
embarrassment (2, 10). High achievers and scale from 1 to 5 ( 1 = Not Severe to 5 = Very
students who have too little time to study for Severe) (14). The results of this study are shown
tests are particularly vulnerable to cheating (11, in Table 2. In several cases (see questions 22-
12). 24) there was almost a full point difference
Students also report that their perception of between the student and faculty perception
faculty reactions to cheating is one of apathy. indicating a lack of communication between
Faculty members often do not report a case of faculty and students. Some of the most
student cheating to the institutional justice problematic areas of disagreement (see questions
system, either for fear of legal repercussions or to 3, 5, 12, 14, and 15) indicate a educational moral
prevent hurting the reputation of the student. laxity on the part of the students.
Instead, many faculty members prefer to handle One may interpret these results in two
each case on an individual basis, sending a signal different ways. One possibility is that the results
to students that the repercussions for cheating are reflect stricter standards developed by faculty
minimal (6, 13). This signal is tantamount to members as they moved in their careers. In other
acceptance of academic dishonesty as a fact in words, their perception reflects a more mature
higher education by both faculty and students. evaluation of the scenario being considered. If
An added problem is that faculty and this interpretation is correct, one also would
students often do not agree on what actions expect students to improve their moral standards
constitute cheating in and out of the classroom as they mature. In other words, the students’
(14-17). The literature recommends that college perception of what constitutes misconduct,
teachers should be very specific in their should not have any influence in their future
definition of academic dishonesty, giving professional conduct. This hypothesis, however,
concrete examples, and then following through does not take into consideration that the faculty
on consistent discipline when cheating occurs members polled in this study already had a
(18, 19). In an effort to determine the level of different perception of what constituted cheating
50
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Turrens, et al., Undergraduate academic cheating

Question Faculty Student P

1 Looking directly at another persons paper to copy 4.88 ± 0.67 4.38 ±1.29 0.0017
an answer during a test
2 Using "crib notes" or a "cheat sheet" during a test 4.83 ± 0.70 4.32 ± 1.31 0.0016
or class assignment
3 Getting a copy of the test prior to taking it 4.80 ± 0.76 3.94 ± 1.23 0.0001
4 Having/paying someone to do homework or at- 4.76 ± 0.81 4.06 ± 1.20 0.0001
home projects for you
5 Copying someone’s homework 4.63 ± 0.87 3.77 ± 1.19 0.0001
6 Using answer book or keys to get homework 3.95 ± 1.24 3.10 ± 1.34 0.0001
answers
7 Leaving the test to go to the restroom/or another 4.77 ± 0.84 4.24 ± 1.33 0.0022
place to get answers
8 Answering "here" or signing someone’s name 4.71 ± 0.79 3.55 ± 1.32 0.0001
when he/she is absent
9 Copying someone’s paper to work and putting your 4.82 ± 0.73 4.17 ± 1.30 0.0001
name on it
10 Trying to influence a teacher to give you a better 3.46 ± 1.31 2.83 ± 1.30 0.0011
grade
11 Using sorority/fraternity test files 3.56 ± 1.46 3.05 ± 1.47 0.0178
12 Finding someone’s idea and using it as your own 4.36 ± 1.00 3.77 ± 1.32 0.0009
13 Asking for answers with gestures or sign language 4.54 ± 1.01 3.93 ± 1.39 0.0010
during an in-class assignment
14 Plagiarism of resource materials or documented 4.76 ± 0.75 4.06 ± 1.39 0.0010
work
15 Using another’s research for your own benefit 4.31 ± 1.13 3.67 ± 1.40 0.0008
16 Watching someone cheat without reporting it 3.51 ± 1.23 2.88 ± 1.26 0.0007
17 Not carrying your weight in a group project for 3.93 ± 1.17 3.62 ± 1.36 0.0991
which everyone gets the same grade
18 Using sources on homework which the professor 4.15 ± 1.16 3.59 ± 1.26 0.0526
told you not to use
19 Getting a teacher’s copy of a test to sell 4.62 ± 1.03 4.22 ± 1.31 0.0072
20 Conducting group sessions to swap or check the 2.71 ± 1.35 2.15 ± 1.34 0.0166
accuracy of answers
21 Giving answers with gestures or sign language 4.50 ± 1.18 3.83 ± 1.30 0.0017
during an in-class assignment
22 Lying to a teacher about why you are not prepared 4.22 ± 1.98 3.27 ± 1.31 0.0000
in class
23 Taking money for doing someone’s work 4.58 ± 1.01 3.62 ± 1.33 0.0001
24 Glancing at another paper and seeing something to 4.40 ± 1.15 3.49 ± 1.24 0.0000
jar your memory
25 Working with someone else on a take-home exam 3.92 ± 1.37 3.06 ± 1.37 0.0004
Table 2. Perception by Faculty and Students of Cheating Behavior in College. 140 students and 108 faculty members were
asked to assign a value to the perceived severity of the behavior on a scale of 1 to 5, with 5 being most severe. The results
are presented as average ± SD. The study was carried out at the University of Montevallo during the Fall of 1997.

when they were in college. They grew up with a history, they will always develop the correct
different set of standards, in an environment in moral values as they become professionals.
which cheating was not as prevalent. Thus, An alternative hypothesis is that, although
accepting this hypothesis would imply that the moral standards of most individuals increase
regardless of the predominant moral values through life, some of these individuals do not see
among college students at any given point in any need to change their values. For them the
51
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
concept of “misconduct” disappears. The school and college. Thus we see a moral laxity
concern of those interested in maintaining high beginning at the high school level (or before) and
post-secondary educational standards is that the progressing, probably with more cheating
habits established by some college students will occurring rather than less, as the level of the
continue to be their habits in graduate school, academic workload increases.
employment and research in the future. One of the established patterns of human
Therefore, an increase in the proportion of an development is the relative stability of
undergraduate students involved in academic personality traits and behavioral habits over the
misconduct is likely lead into an increased life span. Thus, traits of dishonesty in the face of
incidence of professional misconduct in the hard or demanding intellectual work in college,
future. will, in all likelihood, remain stable
The current situation is likely to deteriorate characteristics as these college students grow
even more. The development of the Internet at older. One cognitive/moral development
the end of the 20th century has also increased the theorist, Kohlberg, proposed a universal set of
number of cheating episodes by providing tools discrete stages of moral development based on
that were not available even 10 years ago. Piaget’s theory of cognitive development (21,
Students may now download an enormous 22). As a child develops more complex and
amount of information in seconds, which may be different modes of thinking and reasoning, the
incorporated into a paper with a couple of child should also be able to make more complex
keystrokes. Moreover, several virtual companies and adaptive moral judgments. Kohlberg
have proliferated offering term papers in all proposed a six-level moral developmental
disciplines on a per page cost (see for example, sequence. At Levels 1 and 2, there is a basic
www.schoolsucks.com, www.ezwrite.com, desire to escape punishment and to win some
www.academictermpapers.com, etc.). In the last level of approval from significant others. At
two years there has been a increase in number of Levels 3, 4, 5, and 6, the individual may progress
cases of plagiarism by students who simply from living up to others’ expectations, to
download text from the internet, not just at the following rules to maintain the social order and
University of South Alabama and the University avoid chaos, to adhering to a social contract only
of Montevallo but also at many other institutions. when it appears to be valid to the individual, and,
When confronted by the faculty, these students finally, to upholding moral judgments and
are dismayed at getting caught, but many will principles despite potential harm or threat to
repeat similar behaviors in the future. The only oneself because of their intrinsic worthiness.
tools available to faculty to identify these cases is Kohlberg proposes that rarely do most
to search the web for a specific (unique) individuals progress in moral development past
paragraph in the paper or to contract the services Level 3 or perhaps 4 (21, 22). We do the “right”
of commercial search engines (for example, thing in any given situation to garner favor and
www.plagiarism.org) that can look for the papers approval from others who expect a substantial
sold to students by Internet companies. The first effort from us. And, if we perceive the rules that
procedure is time-consuming and limited. Hiring are in place for us to follow to be unfair or
the services of a company to track these papers nonsensical, we may make a judgment to avoid
down still requires someone to enter the text in complying with those rules on what we call
the Internet and also the becomes too expensive. moral grounds.
Since the formative years of college are With Kohlberg’s postulations in mind, it is
important in setting many of our standards, as the then easy to hypothesize that an individual who
students’ academic standards decrease future learned to cheat in academic situations without
professionals may find it easier to engage in active reprisal from faculty or a school
scientific misconduct as they will perceive it to administration, would tend to repeat those
be less immoral and more expedient. For cheating behaviors in future learning/academic/
example, a study done with 2,459 sophomore research situations as a way to gain approval for
medical students showed that 4.7% admitted to completion of the assignment or project. In
cheating while 66.5% admitted to having heard addition, if the adult who participated in
of cheating among their peers (20). About 70% academic dishonesty all the way through
of the students that admitted having cheated in graduate school may view the demands of a
medical school also admitted to cheating in high thesis or dissertation committee as non-valid, that
52
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Turrens, et al., Undergraduate academic cheating
individual may engage in academic dishonesty Journal of College Student Development 1989; 30: 401-
with an almost-clear conscience. The 406.
requirements of “publish or perish,” then, in the 14. Fleming KD, Keegan DM, Staik IM,, Burling JW.
post-academic world may become “non-valid” in Faculty perception of cheating in college: Toward a
campus consensus. 1998; Paper presented at the 1998
the professional’s mind, and the individual may NCUR (National Conferences on Undergraduate
continue to participate in dishonesty in research. Research) meeting, Salisbury State University, MD.
In summary, the correlation between cheating 15. Livosky M, Tauber RT. Views of cheating among
in high school, college and in medical school college students and faculty. Psychology In The
supports our hypothesis that future professional Schools 1994; 31: 72-82.
misconduct will also show a positive correlation 16. Roig M, Ballew C. Attitudes toward cheating by college
with previous history. Thus, we propose that students and professors. 1992; ERIC NO: ED349895.
part of the efforts to promote integrity among 17. Falleur D. An investigation of academic dishonesty in
future professionals should be devoted to curbing Allied Health. Journal of Allied Health 1990; 19:313-
324.
cheating at the undergraduate level since an 18. Davis SF, Ludvigson HW. Additional data on academic
increase in one is likely to increase the other. dishonesty and a proposal for remediation. Teaching of
Psychology 1995; 22: 119-121.
Bibliography 19. Beemsterboer PL. Academic integrity: What kind of
1. Aaron RM, Georgia RT. Administrator perceptions of students are we getting and how do we handle them
student academic dishonesty in collegiate institutions. once we get them? Journal of Dental Education 1997;
NASPA Journal 1994; 31: 83-91. 61: 686-688.
2. Diekhoff GM, LaBeff EE, Clark RE, Williams LE, 20. Baldwin DC Jr., Daugherty SR; Rowley BD, Schwarz
Francis B, HainesV J. College cheating: Ten years later. MR. Cheating in medical school: a survey of second-
Research in Higher Education 1996; 37: 487-502. year students at 31 schools. Acad. Med. 1996; 71: 267-
3. Haines VJ., Diekhoff GM, LaBeff EE, Clark RE. 273.
College cheating: Immaturity, lack of commitment, and 21. Kohlberg L Essays on moral development: Vol. II. The
the neutralizing attitude. Research in Higher Education psychology of moral development. San Francisco:
1986; 25: 342-354. Harper & Row; 1984.
4. Holm TT. Ignorance of the crime is no excuse: Do 22. Kohlberg L. The development of moral judgment and
public speaking students even know what cheating is? moral action. In L. Kohlberg (ed.), Child psychology
Paper presented at the 84th Annual Meeting of the and childhood education: A cognitive-developmental
National Communication Association (New York, NY, view. New York: Longman; 1995, pp. 259-328.
November 21-24, 1998). ERIC NO: ED427355.
5. Lord T, Chiodo D. A look at student cheating in college
science classes. Journal of Science Education and
Technology 1995; 4: 317-324.
6. Maramark S, Maline MB. Academic dishonesty among
college students. Issues In Education. 1993; U.S.
Department of Education, Office of Research, p. 5,
ERIC NO: ED360903.
7. Whitley BE Jr. Factors associated with cheating among
college students: A review. Research in Higher
Education 1998; 39: 235-274.
8. Graham MA, Monday J, O’Brien K, Steffen S. Cheating
at small colleges: An examination of student and faculty
attitudes and behaviors. Journal of College Student
Development 1994; 35: 255-260.
9. Genereux RL, McLeod BA Circumstances surrounding
cheating: A questionnaire study of college students.
Research in Higher Education 1995; 36: 687-704.
10. Moffatt M. Undergraduate cheating. 1990; ERIC NO:
ED334921.
11. Greene AS, Saxe L. Everybody (else) does it: Academic
cheating. 1992; Paper presented at the Annual Meeting
of the Eastern Psychological Association (Boston, MA,
April 3-5), ERIC NO: ED347931.
12. Meade J. Cheating: Is academic dishonesty par for the
course? Prism 1992; 1: 30-32.
13. Jendrek MP. Faculty reactions to academic dishonesty.
53
2. Institutions and Professions
Comprehensive Guidelines for the Responsible Conduct of Researchers
Gregory Brock, Department of Family Studies, University of Kentucky, USA
Sandra Sutter, Department of Family Studies, University of Kentucky, USA
Ada Sue Selwitz, Office on Research Integrity, University of Kentucky, USA

Keywords: Conduct guidelines, Research ethics, Researcher ethics

In 1989, the Department of Health and Human Services (DHHS) through the Public Health Service
defined research misconduct and established regulations for reporting scientific misconduct among
awardee and applicant institutions (1). The focus of this regulation was on fabrication, falsification,
and plagiarism. More recently DHHS has shifted emphasis toward preventing misconduct and to the
promotion of Responsible Conduct in Research (RCR).
Success in implementing regulatory initiatives on research integrity has been stymied by several
factors. There is disagreement about the extent of research misconduct. Steneck (2) reported that
fewer than 200 cases of misconduct have been documented by federal government research
investigation offices over the past 20 years. Indirect evidence also cited by Steneck, however,
suggests that misconduct may occur far more frequently.
Additionally, there is a lack of clarity about what amounts to research misconduct. In 1989, the
term focused on, “…fabrication, falsification, plagiarism, or other practices that seriously deviate
from those that are commonly accepted within the scientific community for proposing, conducting, or
reporting research.”(1). Defining deviant practice as well as what is common practice is particularly
challenging in view of the rapid development now occurring within many scientific disciplines—what
was deviant can become common practice. Plus, collaboration among academic disciplines, between
universities and industry, between universities and government, and between international research
teams creates new syntheses that further complicate our understanding of what constitutes common
practice. In an effort to address these issues, regulators have turned to requiring training of
researchers as one means of communicating that the incidence of misconduct is troubling. Training
objectives also clarify what amounts to misconduct.
On December 1, 2000, the DHHS Office of Research Integrity adopted and published the final
PHS Policy on Instruction in the Responsible Conduct of Research that delineates RCR training
requirements to all research investigators applying for or using PHS funds and their institutions (3).
Although nine core areas of instruction are specified, the policy does not establish the exact content in
the form of standards and principles within each area. In complying with this mandate, each
institution will be responsible for its own content.
Much attention in the RCR literature has been directed to standards within specific areas, such as
authorship, peer review, and collaborative practices. Presentations at national conferences and

Corresponding author: Gregory Brock, Ph.D., Department of Family Studies, University of Kentucky, 316 Funkhouser
Building, Lexington, KY 40506-0054, 859-257-7742 (voice), 859-257-321 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
institutional committees have addressed RCR associations. Guidelines were then sorted into
practice standards. As well, many professional discrete thematic categories. These categories
associations have established standards of were called principles because they identified
conduct within their ethical codes. Institutional core values of research practice. Three groups of
policies such as Guidelines for the Conduct of principles emerged from the analysis: General,
Research at the National Institute of Health have Professional, and Focused. Subprinciples also
also incorporated a selection of RCR topics (4). were defined that served to elucidate
However, no single set of principles contemporary issues rather than merely
encompassing all aspects of responsible conduct exemplifying situations in which the principles
of research exists in unified form. might apply. A series of revisions were made
Grinnell (5) pointed out that “…promoting after obtaining feedback from research
responsible conduct of science requires a clear colleagues and university administrators.
description of what doing science entails.” In What emerged was a comprehensive set of
addressing why standards are important, Frankel guidelines for the conduct of researchers more
(6) discussed the need of the general public for akin to a code of conduct for a profession (see
accountability in science, and how a set of attached guidelines). These guidelines provide a
standards not only meets this need but also broad-based foundation for the safe and effective
increases trust in the scientific community. practice of research across disciplines, settings,
Frankel noted specific benefits to establishing methods, and questions. Our intent in presenting
ethical standards: Standards provide an enabling them here is to increase the awareness and
document, professional socialization, public sensitivity of institutional policy makers to the
accountability, gain public trust/support, improve many issues that researchers must attend to in the
public relations, self-preservation, deterrence, conduct of their professional responsibilities. By
professional support, and are a source of public presenting the results of our analysis, we wish to
policy. Standards also provide guidance when an further the discussion about the content of RCR
ethical course of action is unclear. Mastroianni training.
and Kahn (7) point out that training students in
the basics of RCR is crucial to the continued Acknowledgements:
maintenance of public trust in the scientific Support was provided by the Vice President for
community by cultivating the integrity of Research and Graduate Studies, University of
research practices. However, results on the Kentucky.
effectiveness of RCR training thus far are
inconclusive (8, 9). Brown and Kalichman (9) References
offer the interpretation that a lack of consensus 1. Department of Health and Human Services, Public
on what constitutes misconduct may contribute to Health Service. Responsibilities of Awardee and
the lack of clarity on the effectiveness of training. Applicant Institutions for Dealing With and Reporting
Frankel (10) advocates the development of Possible Misconduct in Science. Federal Register Vol.
research standards as the single most important 54 No. 151 Tuesday, August 8, 1989.
2. Steneck, N. Assessing the Integrity of Publicly Funded
step in promoting scientific integrity and
Research: A Background Report. Paper prepared for the
handling misconduct. Faced with the new Department of Health and Human Services, Office on
training requirements established by the PHS, Research Integrity Research Conference on Research
this step is particularly important for promoting Integrity on November 2000, Bethesda, MD.
and supporting a climate of integrity at the 3. PHS Policy on Instruction in the Responsible Conduct
organizational level that can function in a of Research, Department of Health and Human
reciprocal fashion to influence and be influenced Services, Public Health Service, Office of Research
by individual actions. Integrity. Available at URL: http://ori.dhhs.gov/html/
Initially, the purpose of the document programs/rcrcontents.asp
4. American Association of Medical Colleges. Developing
presented here was to provide a comprehensive
a code of ethics in research: A guide for scientific
set of guiding principles to serve as a basis for societies, executive summary. Conference Materials of
RCR training at the University of Kentucky. the Public Responsibility in Medicine and Research;
Content analysis was applied to an exhaustive list 1997 May 13-14; Bethesda, Maryland.
of behavioral guidelines identified in a thorough 5. Grinnell, F. Ambiguity, trust, and the responsible
review of the research integrity literature conduct of research. Science and Engineering Ethics,
including ethics codes of professional 1999; 5 (2): 205-214.

58
––––––––––––––––––––––––––––––––––– Brock et al., Comprehensive Guidelines for the Responsible Conduct of Researchers
6. Frankel, M. S. Developing ethical standards for a survey. Science and Engineering Ethics 1996; 2: 89-
responsible research: Why? Form? Functions? Process? 114.
Outcomes?, Journal of Dental Research 1996; 75 (2): 9. Brown & Kalichman. Effects of training in RCR: A
832-835. survey of graduate students in experimental sciences.
7. Mastroianni, A. C., & Kahn, J. P. The importance of Science and Engineering Ethics 1998; 4: 487-498.
expanding current training in responsible conduct of 10. Frankel, M. S. Scientific community must set the
research. Academic Medicine 1998; 13 (12): 1249-1254. standards. Forum for Applied Research and Public
8. Eastwood, S., Derish, P., Leash, E., & Ordway, S. Policy, Spring 1998
Ethical issues in biomedical research: perceptions and
practices of postdoctoral research fellows responding to

Guidelines for the Responsible Conduct of Researchers

Preamble: Advancing the scientific record is the noble task of those who conduct research. In large
part the quality of that record is the product of inquiry. Ranging well beyond the conduct of research
however is the realm of activities constituting the work of researchers that influences the public trust,
that affects global well-being, and that indirectly affects the scientific record. The guidelines
presented here define expectations so that researchers uphold the highest ethical standards by
practicing within the bounds of both effectiveness and safety.
Important, sustaining values that support humankind and global well-being serve as the basis for
three groups of principles and sub-principles. (1) General principles apply to all research contexts.
(2) Professional principles define relations among researchers and practices that constitute the
scientific method. (3) Focused-principles address discrete aspects of research practice for particular
investigations, research contexts, or scientific disciplines. Sub-principles elucidate contemporary
issues rather than identifying the component issues of any principle.
Where governmental laws contradict these guidelines, researchers are cautioned to seek
consultation from appropriate authorities and colleagues. Resolution is not always possible,
consequently, researchers act so as to benefit the greater good even if that path demands personal
sacrifices.
In an effort to create a research climate worthy of the public trust, it is incumbent upon
researchers to report any breech of these guidelines to an appropriate authority. Where there is no
relevant authority, researchers are obliged to focus public media attention on wrong doing.
These guidelines apply to professional and amateur researchers, students, research technicians,
research administrators, as well as private, public, and governmental research agency personnel.

General Principles

General Principle 1: Commitment to Society and to Global Well-being

Researchers protect the interests of society within a broader commitment to global well-being. They
recognize that the public has entrusted them to uphold the integrity of the scientific record.
1.1 Researchers do not obligate themselves to withhold research findings that may jeopardize the
health or well-being of others.
1.2 Researchers take active steps to prevent the misuse of their findings that may jeopardize the
well-being of others.
1.3 Researchers take active steps to correct errors or oversights in proposing, conducting, or
reporting research.
1.4 Researchers present themselves to the public in a competent, sincere, and trustworthy man-
ner.
59
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

General Principle 2: Commitment to Competency

Researchers are aware they are responsible for maintaining professional competency and remaining
knowledgeable within their areas of expertise.
2.1 Researchers conduct their work within the scope of their own training and knowledge base.
2.2 Researchers recognize they are vulnerable to stress and impairment. When stress or impair-
ment interferes with their ability to conduct professional responsibilities, researchers seek
assistance.
2.3 Researchers ensure that all persons who assist in the conduct of their research are adequately
trained and perform their responsibilities competently.
2.4 Researchers inform their work with views, values, and co-workers from diverse sources.
2.5 Researchers foster a scientific community in which discrimination based on gender, race age,
sexual orientation, religious affiliation, ethnic or national origin does not occur.

General Principle 3: Understanding Laws, Regulations, and Policies

Researchers are aware of and stay informed of professional, institutional, and governmental
regulations and policies in proposing, conducting, and reporting research.
3.1 Researchers take active steps to resolve discrepancies when policies or regulations are
unclear or contradict one another.

General Principle 4: Conflicts of Interests

Researchers are cognizant that conflicts of interest occur in the context of professional activities and
they recognize and avoid them.
4.1 When researchers cannot avoid real or perceived conflicts of interest, they seek consultation
and take active steps to minimize bias, flawed judgment, harm, or exploitation.

Professional Principles

Professional Principle 5: Peer Review

Researchers respect others’ rights to have work reviewed in a confidential, timely, and objective
manner.
5.1 Researchers assess and disclose multiple roles or allegiances which may undermine the
confidential and fair review of others’ work.
5.2 Researchers take active steps to protect the integrity of review materials and guard the
intellectual property of others.

Professional Principle 6: Research Management and Data Access

Researchers clearly and authentically record data and methods. They protect the integrity of their
research materials. They make data, methods, and materials available to others for analysis or
replication.
6.1 Researchers select materials appropriate for data acquisition, recording, and storage.
6.2 Researchers stay informed of and implement policies for appropriate storage and disposal of
research materials.
6.3 Researchers take active steps to select methods and materials that protect research partici-
pants’ right to privacy.
6.4 Researchers take active steps to safeguard data when using electronic or Internet-based
methods.
6.5 Researchers are cognizant of the ownership of their research data, methods, and findings.

Professional Principle 7: Commitment to Credibility

Researchers engage in practices that are currently accepted within the scientific community to
propose, conduct, and report research.
60
––––––––––––––––––––––––––––––––––– Brock et al., Comprehensive Guidelines for the Responsible Conduct of Researchers

7.1 Researchers practice honest stewardship of their research resources and use recognized
accounting methods.
7.2 Researchers do not conduct their professional responsibilities in a manner that is intentionally
deceitful or with reckless disregard for the truth.
7.3 Researchers who witness or suspect fraud or misconduct follow established procedures to
preserve the integrity of the scientific record.
7.4 Researchers accused of fraud or misconduct do not harass those believed or known to have
made accusations against them.
7.5 Researchers do not misrepresent their work by omitting data that changes the meaning or
significance of their findings.
7.6 Researchers do not fabricate or falsify data.
7.7 Researchers do not present or publish component findings of a larger body of work if misun-
derstanding may result or to conceal findings.

Professional Principle 8: Mentoring, Training, and Supervisory Relationships

Researchers nurture the intellectual, technical, ethical, and career development of their trainees,
supervisees, and students.
8.1 Researchers recognize that trainees, supervisees, and students have needs unique to their
individual strengths and limitations. Researchers provide guidance, constructive feedback,
and assistance that matches the changing needs of each trainee, supervisee, or student.
8.2 Researchers establish clear and appropriate rules and boundaries in their relationships with
trainees, supervisees, and students.
8.3 Researchers do not engage in sexual harassment, disrespect the character of, or impede the
progress of their trainees, supervisees, and students.
8.4 Researchers recognize that exploitation is a risk in relationships where differences in power
exist. They avoid conflicts of interest and dual relationships. Sexual interaction with subordi-
nates is avoided.
8.5 Researchers take active steps to inform trainees, supervisees and students of supervisors’
responsibilities to avoid dual relationships.

Professional Principle 9: Authorship and Publication Practices

Researchers respect the intellectual property rights of others.
9.1 Researchers attribute credit for others’ words and/or ideas in proposing, conducting, or
reporting their own work.
9.2 Researchers facilitate discussion and set ground rules early in collaborative relationships
regarding authorship assignment.
9.3 Researchers assume responsibility for the accuracy of research reports for which they claim
full or co-authorship.
9.4 Researchers preserve the integrity of the scientific record by taking active steps to correct
errors in the publication of their findings.
9.5 Researchers do not submit or publish previously published materials without appropriate
citation.
9.6 Researchers respect the privacy of others’ unpublished work.

Professional Principle 10: Responsibilities to Colleagues and Peers

Researchers recognize they are members of the scientific community and respect the contributions of
others to the scientific record.
10.1 Researchers clarify early in a collaborative project the expectations and responsibilities
among those involved.

61
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

10.2 Researchers do not impede the progress of others’ work.

10.3 Researchers protect the integrity of intellectual property and research materials when
reviewing others’ work.
10.4 Researchers take active steps to maintain positive relations among team members and to
seek consultation if necessary to resolve interpersonal conflicts.

Focused Principles

Focused Principle 11: Protection of Human Participants

Researchers respect the dignity of human participants and take active steps to protect their well-
being. They follow institutional, professional association, and governmental ethical and
regulatory guidelines.
11.1 Researchers ensure that each participant gives voluntary and informed consent regardless of
age, race, gender, ethnic or national origin, sexual orientation, mental or physical health
status, or incarceration.
11.2 Researchers take active steps to evaluate and to minimize potential risks to participants.
11.3 Researchers respect each participant’s right to privacy, and they take active steps to protect
confidentiality of data or other disclosures.
11.4 Researchers take active steps to achieve an equitable balance of benefits and risks to each
participant.
11.5 Researchers honor fairness and equity in the selection of research participants.

Focused Principle 12: Care and Use of Animals for Research

Researchers are stewards of animals used for research. They follow institutional, professional
association, and governmental ethical and regulatory guidelines.
12.1 Researchers substitute inanimate materials and processes for animals where appropriate.
When this is not possible, researchers make active efforts to use species that may be less
susceptible to pain and distress.
12.2 Researchers take active steps to use procedures which reduce the incidence and/or severity
of pain and distress experienced by animals.
12.3 Researchers take active steps to reduce the use of animals to the minimum number neces-
sary to yield valid answers to their research questions.

Focused Principle 13: Commitment to Native Populations and Other Identifiable Groups
Researchers respect the rights and protect the interests of Native populations and other
identifiable groups.
13.1 Researchers who work with Native populations and other identifiable groups recognize that
to minimize risks and to maximize benefits to individuals and to populations themselves
there is value in obtaining the advice, participation, and viewpoints of those individuals and
populations in formulating research questions, designing research methods, collecting and
analyzing data, and in reporting results.
13.2 Researchers recognize that consent from or consultation with group authorities or represen-
tatives is sometimes necessary before obtaining consent from individuals within Native
populations or other identifiable groups.
13.3 Researchers take active steps to distinguish individual property both tangible and intangible
from collective property owned by Native populations or other identifiable groups.
13.4 Researchers take active steps to reduce the risk to Native populations or other identifiable
groups that result from misuse of their research findings.

62
––––––––––––––––––––––––––––––––––– Brock et al., Comprehensive Guidelines for the Responsible Conduct of Researchers

Focused Principle 14: Genetic Research and Technology

Researchers strive to preserve and protect global well-being from the unintended consequences of
genetic research.
14.1 Researchers involved in genetic research take active steps to identify potential risks and
benefits to research participants. They inform participants of the possibility that risks may
not yet be identified.
14.2 Researchers take active steps to protect the confidentiality of genetic materials collected
from human participants and do not allow the use of these materials for purposes which
may discriminate against or harm an individual or group of individuals.
14.3 Researchers are sensitive to social, physical, psychological and environmental factors that
may influence individuals’ consent to participate in genetic research.
14.4 Researchers inform individuals, their families, and Native
and other identifiable populations of the disruptive influence that genetic research may have on
their lives. They take active steps to minimize disruptions.
14.5 Researchers are cognizant of the increasing complexity of the ethical concerns about genetic
research. They stay informed of the developing research guidelines as well as the public
discourse about genetic research.
14.6 Researchers actively participate in the development and refinement of ethical standards in
this area.

63
Research Integrity in Social Work: Status, Issues, and Future Directions
Margaret Gibelman, Wurzweiler School of Social Work, Yeshiva University, USA

Keywords: Scientific integrity, social work, code of ethics, scientific misconduct, institutional review boards

This paper explores the issue of scientific integrity in social work and its implications for the training
of social work researchers. Data concerning a growing body of cases in which allegations have been
made and/or violation of legal and ethical research standards have been substantiated illustrate that
the integrity of research in social work and related fields is a growing concern. However,
mechanisms to review and monitor social work research are under-developed compared to other
disciplines. A research agenda is offered to assess the status of institutional systems to review and
monitor research in social work and, concurrently, determine social workers’ familiarity with the
profession’s ethical code as it relates to research integrity. Implications for faculty and practitioner
education and training and the development and enforcement of systems to review the integrity of
research protocols are explored.
Scientific misconduct or, more positively, appropriate conduct in the realm of research inquiry, is
a topic that has received very little attention in the social work literature. Unfortunately, this is
because social workers have not, historically, been strong contenders in the successful competition for
federal research grants, particularly large-scale research protocols (1, 2, 3, 4). Social work research is
still in its infancy compared to research in other disciplines. However, there is a professional
commitment to increase the capacity and productivity of social work research, as evidenced by the
burgeoning number of social work research centers and a growing empirical social work literature
base. This expansion of social work research is not without risks. Although the majority of publicized
cases of scientific misconduct have centered largely on bio-medical research and the applied sciences,
the circumstances associated with these cases have strong implications for the preparation of students
and the standards to which social work researchers will be held. The growing number of cases in
fields related to social work, as discussed below, highlight areas of potential vulnerability.

The Status of Social Work Research

Unlike most of the social and behavioral sciences, social work is a practice-based profession rather
than an academic discipline or field. Social work has been defined as the “applied science of helping
people achieve an effective level of psychosocial functioning and effecting societal changes to
enhance the well-being of all people” (5). Historically, its knowledge base has been predicated upon a
liberal arts perspective and has drawn from psychology, psychiatry, sociology, political science,
economics, and other disciplines to formulate applied practice principles. However, within the past
two decades, social work has striven to define its own unique body of knowledge, an effort
incorporated into the purposes of social work itself, one of which is “the development and testing of
professional knowledge and skills...” (6).
Corresponding author: Margaret Gibelman, Yeshiva University, Wurzweiler School of Social Work, 2495 Amsterdam
Avenue, New York, NY 10033, 212-960-0840 (voice), 212-960-0822 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Although research has always had a place major research centers and the increased
within the purposes of the profession, the larger productivity of social work researchers has been
socio-political environment has, in recent years, long in coming. The mandate to create a
profoundly affected the priority afforded to coordinated research infrastructure had been
research. There is a growing mandate for all echoed for two decades (16, 17, 18, 19).
social workers to incorporate research into their National Institute of Mental Health funding has
practice, a phenomenon underscored by the been a major impetus to establish social work
demands of funding bodies, oversight agencies, research centers at academic institutions. In this
and consumer choice movements for hard data process, however, the profession faces a host of
documenting that programs of service lead to issues and challenges, foremost among them the
tangible results. A leading force has been that of preparation of future researchers, including
managed care, which has brought with it socialization to the ethos of scientific integrity.
heightened demands for accountability, with
particular emphasis on documenting the Ethical Guidelines
successful outcomes of service (7). The latest revision of the Code of Ethics of the
At the same time that external demands to National Association of Social Workers
provide empirical evidence of the impact and (NASW)(20) emphasizes the central role of
outcomes of services grow, social workers, to research: “social workers should contribute to
better protect the interests and well-being of the the knowledge base of social work and share
people they serve, are seeking to empirically with colleagues their knowledge related to
examine the consequences of the managed care practice, research, and ethics. Social workers
movement, itself. This has translated to a concern should seek to contribute to the profession’s
about documenting the effects of managed care literature and to share their knowledge at
(e.g., short-term hospitalization; short-term professional meetings and conferences” (Section
treatment; limited provider choice). These 5.01(d), p. 24). Section 5.02 (b) of the Code
developments have led to the need for a new or (1996) encourages social workers to “promote
enhanced repetoire of research skills on the part and facilitate evaluation and research to
of not only academics and researchers, but contribute to the development of knowledge”
among the totality of social workers directly (p. 25).
providing, supervising, or managing the delivery The Code of Ethics not only seeks to
of human services. establish an obligation on the part of social
The long and ongoing admonishment that the workers to engage in knowledge building
profession must develop an internal research through empirical research, but also provides the
capacity has borne fruit. In fact, a notable basic guidelines for how such research is to be
number of studies have been conducted on the conducted. Specific provisions pertain to risk-
status of research productivity and the scholarly benefit analysis, voluntary and written informed
contributions of social workers (8, 9, 10, 11, 12, consent, protection from harm, confidentiality,
13). Perhaps the most significant influence, and accurate reporting of findings. Further, the
however, on the growing social work research Code sets forth the obligation of social workers
enterprise has been the shift in criteria for tenure to educate themselves and for programs of social
and promotion within academia, which remains work education to provide relevant education
the richest source of social work research (14, concerning responsible research practices.
15). Longevity of academic careers now rests An important caveat about ethical guidelines
firmly on scholarly productivity and standards exists that is idiosyncratic to the profession —
related to both quality and quantity continue to the limited application of the Code to social
rise as social work is increasingly held to the workers. The Bureau of Labor Statistics (21)
same standards as other academic and estimates that there are approximately 542,000
professional units within the university (4). A professional educated social workers in the
related factor in the emphasis on research United States (at the bachelor’s, master’s and
productivity is the growing sophistication of doctoral levels). At the same time, current
faculty in identifying funding sources and membership of the National Association of
competing successfully for publicly supported Social Workers is approximately 155,000. The
research dollars. Code of Ethics is a product of the National
The emergence of schools of social work as Association of Social Workers and, upon joining,
66
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gibelman, Research Integrity in Social Work
members must pledge to abide by the Code. But ethics and that there had not been an explicit
what about the more than 387,000 social workers violation of the Code of Ethics (24). However,
who are not members of NASW and not as social workers increasingly compete
committed to abiding by the provisions of the successfully for federal research funds, they
Code? These social workers may belong to other become subject to the same level of scrutiny as
professional associations which have their own researchers in other disciplines. Similarly, as
ethical guidelines, but data to support this IRBs extend their purview to include privately
contention are lacking (22). Social work supported research, more diligent reviews of
researchers based in institutions of higher social work research protocols can be expected.
education may have their own review and As the social work profession seeks to
oversight procedures, separate from university- enhance its research capability in a credible and
wide IRBs, but again there is an absence of responsible manner, there is much to be learned
substantiating empirical data. An unknown, but from the experience of related disciplines and
impressionistically high proportion of social professions. In recent years there has been a
work research is outside the purview of federal growing number of cases of scientific
funding, which may mean that IRB review misconduct among allied health-related
procedures are not applied. (It should be noted, industries (e.g., nursing, psychology, and
however, that such research is now selectively psychiatry), the predominant theme of which
being reviewed by IRBs to conform with their concerns plagiarism and/or falsification or
own internal procedures, partially reflecting the fabrication of data (25, 26, 27, 28, 29). Eight
prevalence and influence of the growing number cases from the helping professions over the last
of studies sponsored by private sources, decade were identified from media reports, out of
including pharmaceutical companies, in areas an unknown universe of substantiated cases of
such as genetic testing (23).) misconduct. Unlike many cases of misconduct
Finally, in some instances, social work substantiated in the bio-medical fields, these
research may be absent of any oversight by any cases were absent allegations of human subjects
source. This latter scenario is most likely to violations. However, findings of misconduct
prevail among those working in service highlight the diligent reviews to which research
organizations which have not yet established reports are subject and the serious penalties that
review and oversight procedures and may, are levied when ideas are appropriated or results
indeed, not even recognize the need to do so. Of falsified. Sanctions include forced resignations,
particular concern is the mandate for practice criminal prosecution, ineligibility from receiving
agencies to engage in research without publicly supported grants or serving on review
assurances of appropriate procedures and absent panels, and remedial courses in ethics. These
collaborations with educational institutions from sanctions have widespread and serious
which such protocols may be borrowed. implications for how research is conducted and
highlight the potential consequences that may
Learning from the Mistakes of Others ensue when procedural and ethical breaches are
To date, public disclosure of cases of scientific uncovered.
misconduct within the social work research
community have been absent. Over a 10 year Emerging Issues
period of vigilant reporting of scientific The mistakes of researchers of allied disciplines
misconduct, the Chronicle of Higher Education suggest the scope and magnitude of potential
referenced only one situation involving a social areas of scientific misconduct that may similarly
worker. This case concerned a researcher who affect social work. Further, the record on
submitted bogus articles to professional journals misconduct shows that attention to the initial
as part of an experiment to test peer-review review of protocols is only a beginning step in an
practices (24). Because the research did not ongoing process necessary to ensure scientific
involve the use of Federal funds, review of integrity. Although a systematic process for
allegations of ethical misconduct remained reviewing research proposals, including attention
within the purview of the adjudication process of to scientific validity of the study design, can
the NASW. Ultimately, NASW dismissed the alleviate many potential problems, it is in the
complaint, arguing that the issue involved a reporting of research findings, at least to date,
disagreement over research methods rather than that the allegations of scientific misconduct are
67
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
most likely to occur. Reports of research are, in lack of applicability to a large proportion of
fact, reviewed; how research is carried out and social workers. Social work educators, who are
findings reported are subject to scrutiny, and, the major producers of research, are ill-
sometimes, reprisals. This fact presents a represented among the membership of NASW
formidable problem in balancing the traditional and are thus outside of its monitoring and
academic freedom associated with the pursuit of adjudication provisions. Thus, the question of
research and institutional responsibility to ensure what mechanisms govern academic social work
accountability of the outcomes of such research. research remains unanswered.
The extent to which a school of social work can The majority of schools of social work are
monitor the work of its faculty and students is housed in research universities which have their
inherently limited. own IRBs and the logical source of research
While only about 30% of the cases of review and oversight lies with IRBs. However,
scientific misconduct are eventually determined the focus of many, if not most, IRBs on bio-
to be founded, the impact of the allegations is medical research, with the composition of IRBs
profound (30). The investigation of allegations reflecting this emphasis, has limited the informed
consumes significant institutional resources and review of social work protocols. Social and
can ruin careers, even if the allegations are behavioral science research protocols, including
unfounded. If allegations are confirmed, it is those of social work, are often “expedited” and/
lethal to a researcher’s career (see, for example, or are reviewed by researchers who are
31), causes reputational damage to the university, unfamiliar with the nature of such scientific
and may affect public perceptions of the integrity inquiries. (An analogy holds when social and
of all research. Worse, human lives and well- behavioral scientists are asked to participate on
being may be compromised (4). IRBs in the review of bio-medical research.)
Internal systems to prevent and, when Without the procedures in place and a cadre of
necessary, address scientific misconduct are not trained researchers available and able to review
without their critics. There are enormous social work research protocols, social work may
workload implications, particularly for senior well be vulnerable to some of the questionable
faculty who may not have the time or desire to research practices that have been unearthed in
spend their time monitoring junior faculty. There related fields.
are also those who argue that when schools/ The expanding boundaries of what
universities serve as the “scientific validity constitutes scientific integrity are of particular
police” of their own colleagues, they will either relevance to social work researchers. The
join ranks in defense, or, to the other extreme, research conducted by social workers, both
find against their colleagues for fear of students and faculty and agency-based
accusations of institutional bias (32, 33). practitioners, involves interaction with
populations that are often classified as vulnerable
Current Review Mechanisms and confidentiality of data is often an issue.
Since allegations and, in some cases, findings of Direct observations, the administration of
scientific misconduct are, by definition, after-the- questionnaires, review of existing case records,
fact of the activity, the most significant lesson or the introduction of therapeutic interventions
from these cases is the importance of ensuring and the use of control groups that do not receive
that research review and monitoring procedures interventions may be innocuous or, alternatively,
are uniformly followed. The integrity of may pose risks to the emotional, social, or
scientific research is monitored by two main and economic well being of participants (4).
distinct sources: professional associations and Deception, invasion of privacy, lack of informed
their applicable ethical codes and institutional consent, mandatory reporting requirements (such
review boards (IRBs). In social work, these as cases in which potential child abuse is
mechanisms for ensuring research integrity are identified), or the loss of economic benefits (as
less firmly entrenched. As discussed earlier, there may apply, for example, to the disabled or
is no one body with the authority or jurisdiction welfare recipients) are all examples of harm that
to oversee the entirety of the social work research may result from faulty research designs or
enterprise. The guidelines detailed in the misconduct in the implementation of research
profession’s Code of Ethics about ethical protocols (4). Although substantiated cases to
research conduct are, however, limited by their date fall outside of these human protection areas,
68
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gibelman, Research Integrity in Social Work
the nature of the research conducted within the provisions regarding appropriate scientific conduct,
helping professions suggests the potential of such through such means as: (1) an “exit” test of graduating
misconduct. BSW, MSW and doctoral students; (2) a sample
survey of agency practitioners; and (3) a sample
survey of agency administrators charged with
A Call for Research on Social Work responsibility to collect, analyze, and report on client-
Research sensitive data.
Given the relatively undeveloped, but now • An analysis, perhaps through the use of focus groups, of
rapidly expanding research enterprise in social issues and obstacles to the conduct of ethical research
work, there is a clear need for information about which result from the demands of external
how research is monitored and reviewed. The accountability bodies.
number of publicized cases of wrongdoing in • An investigation of the procedures used by schools of
social work to review and monitor faculty and student
fields closely allied with social work suggest that
research, including the scope of such reviews and the
programs of social work education need to extent to which the validity of the science itself is
formulate or revise their procedures for research considered.
review and oversight. Institutional mechanisms • A survey of social work faculty concerning their level of
are needed to ensure that: (1) researchers are participation in university-wide institutional review
cognizant of the ethical issues involved; (2) the boards.
protocols meet university and Federal standards; • A survey of deans and directors of social work
and (3) findings are based on systematic and education programs to identify the frequency, nature,
valid research. The question then becomes and types of issues and problems that have arisen in
regard to studies, once approved and implemented.
whose responsibility it is to monitor such
• A content analysis of material covered in federally
protocols and review the research conducted and prescribed training of researchers and an assessment
how mechanisms can be established which of the applicability of such training to the social and
significantly reduce the potentiality of scientific behavioral sciences.
misconduct. The data emanating from such studies would
Some schools have assembled their own provide a basis for an informed assessment of the
committees to review and pass judgment about extent to which mechanisms for research review
compliance with university and/or federal and monitoring are in place and how well they
research requirements. However, such reviews operate. Such information could form the basis
usually focus on issues of methodology and/or for developing or revising review procedures
informed consent. This is not sufficient given the through university IRBs, through separate IRBs
broadened definition of scientific misconduct, potentially established for the social and behavior
which has been extended beyond the initial focus sciences, or through social work education-
on informed consent, risk levels, and coercion specific structures. Further, such information
(34). The definition of misconduct now includes could be used to develop targeted educational
fabrication, falsification, or plagiarism in programs about research integrity to the social
proposing, performing, or reviewing research, or work community.
in reporting research results” (35, p. 4). The
extent to which social work education programs Conclusion
maintain their own review and oversight Research about social work research has tended
procedures is also unknown. Anecdotal evidence to be descriptive, often focused on admonish-
suggests that such internal program mechanisms ments about the under-developed state of the art
are the exception. Given the limited applicable or analyses of what content areas have been
of the professional Code of Ethics, the unknown researched and what gaps exist. Ethical research
degree of inclusion of social work within the conduct has, by and large, been ignored, in part
purview of IRBs, and the similarly unknown because of the early stage of development of the
degree of school-specific procedures, the need research enterprise. However, the issue of
for “research on the status of social work research integrity takes on increasing importance
research” is suggested. Possible areas of inquiry as social work gains a legitimate role in the
include: conduct of scientific inquiry. The profession is
• An analysis of the social work education curriculum to likely to experience a stronger imperative to
ascertain the degree to which ethical conduct is a
engage in research as demands for accountability
component of research courses.
• An assessment of social workers’ familiarity with ethical
and documentation of the outcomes of human
services continue to grow.
69
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Strategies to ensure research integrity 10. Gelman, S.R., Gibelman, M. A quest for citations?: An
depend, first, on a clearly formulated agenda analysis and commentary on the trend toward multiple
based on an assessment of the current status of authorship. J Soc Work Educ 1999; 35: 203-13.
review and monitoring systems. Based on hard 11. Green, R.G., Baskind, F.R., Best, A.M., Boyd, A.
Getting beyond the productivity gap: assessing variation
data, the professional schools of social work and in social work scholarship. J Soc Work Educ 1997; 33:
their universities can assume the task of 541-53.
modifying and strengthening procedures in a 12. Ligon, J., Thyer, B.A., Dixon, D. Academic affiliations
manner that is reflective of the burgeoning social of those published in social work journals: a
work research enterprise. Means of prevention productivity analysis, 1989-1993. J Soc Work Educ
as well as amelioration need to be developed, 1995; 31: 369-76.
codified, and enforced. In this process, there is a 13. Smith, S., Baker, D., Campbell, M., Cunningham, M.
need to define the parameters of both appropriate An exploration of the factors shaping the scholarly
scientific conduct and what constitutes productivity of social work academicians. J Soc Serv
Res, 1985; 8: 81-99.
misconduct as it relates to social work research 14. Marsh, J.C. Should scholarly productivity be the
and to elaborate on its meaning with some degree primary criterion for tenure decisions? Yes! J Soc Work
of precision. Clear university and school Educ 1992a; 28: 132-4.
standards, widely publicized, and ongoing 15. Marsh, J.C Response to Dean Weick. J Soc Work
education regarding appropriate scientific Educ, 1992b; 28: 138-9.
conduct would help alleviate actual or potential 16. Austin, D. Findings of the NIMH task force on social
problems as social work secures a more extensive work research. Res on Soc Work Pract 1992; 2: 311-
and important role in the production of research. 322.
17. Council on Social Work Education. Research utilization
in social work education. New York: CSWE; 1981.
Bibliography 18. Council on Social Work Education. A planning proposal
1. Austin, D. Research and social work: educational to establish a National Institute for Social Work
paradoxes and possibilities. J Soc Sci Res 1979; 2: 159- Research. New York: CSWE; 1983.
76. 19. Estes, R.J. Social work research centers. Soc Work
2. Fanshel, D. National Conference on the future of social Res Abstr 1979; 15: 3-16.
work research: Final report. Washington, DC: National 20. National Association of Social Workers. Code of ethics.
Association of Social Workers, 1979. Contract # 278- Washington, DC: NASW; 1996.
77-0079[SM]. Sponsored by the National Institute for 21. U.S. Department of Labor, Bureau of Labor Statistics.
Mental Health. Tabulations from the Current Population Survey, 1998
3. Rubin, A., Gibelman, M. Social work research in annual averages. Washington, DC: BLS; 1998.
mental health: the state of the art. Rockville, MD: 22. Gibelman, M., Schervish, P. Who we are: a second
Division of Human Resources/Division of Prevention look. Washington, DC: NASW Press; 1997.
and Special Mental Health Programs, National Institute 23. IRBs and human stem cell research protocols. Hum Res
of Mental Health; 1984. Report no.: 83M055981901D Rep 2000 Jan 15; 3-4.
& 83M052517801D. 24. Social workers’ group plans no disciplinary action
4. Gibelman, M., Gelman, S.R. Learning from the against researcher who submitted bogus articles. Chron
mistakes of others: a look at scientific misconduct High Educ 1989 Apr 15; Sect. A: 5.
within academia. J Soc Work Educ 2001; 37(2): 241- 25. Burd, S. Nursing professor found guilty of scientific
254.. misconduct. Chron High Educ 1995 July 7; Sect. A:23.
5. Barker, R.L. Social work dictionary. 4th ed. 26. Burd, S. Research powerhouse under the microscope.
Washington, DC: NASW Press; 1999. Chron High Educ 1994 Jun 15; Sect. A: 24.
6. Council on Social Work Education. Handbook of 27. Campbell, P.W. Researcher found guilty of misconduct.
accreditation standards and procedures. Alexandria Chron High Educ 1999 Jan 15; Sect. A:34.
(VA): CSWE; 1994. 28. Walker, P.V. Public health service finds 3 scientists
7. Gibelman, M., Whiting, L. Negotiating and contracting guilty of misconduct. Chron High Educ 1996 Jan 12;
in a managed care environment: considerations for Sect. A: 26
practitioners. Health Soc Work 1999; 24: 180-90. 29. Wheeler, D.L. Suggestions offered for investigating
8. Baker, D.R., Wilson, M.V.K. An evaluation of the reports of biomedical-research fraud. Chron High Educ
scholarly productivity of doctoral graduates. J Soc 1991 Nov 27; Sect. A: 6.
Work Educ 1992; 28: 204-12. 30. Goodman, B. Scientists exonerated by ORI report
9. Bloom, M., Klein, W.C. Publications and citations: a lingering wounds. Scientist 1997, June 9; 11: 124.
study of faculty at leading schools of social work. J Soc 31. Schmidt, P. U. of Arizona fires prominent researcher
Work Edu 1995; 31: 377-87. over charges of scientific misconduct. Chron High Educ
1998 Aug 14; Sect. A: 12.

70
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gibelman, Research Integrity in Social Work
32. Block, L.E. Science’s credibility depends on effective
peer review. Chron High Educ 1991 July 31; Sect. B:2.
33. Silbergeld, E.K. Annotation: Protection of the public
interest, allegations of scientific misconduct, and the
Needleman case. Am J Public Health 1995; 85: 165-6.
34. Broad, W.J. Data tying cancer to electric power found
to be false. New York Times 1999 July 24; Sect. A:1,
A10.
35. New definition proposed for research misconduct. Hum
Res Rep 1999 Dec 14; 4.

71
Organizational Influences on Scientific Integrity
Michael D. Mumford, Department of Psychology, University of Oklahoma, USA

Whitney B. Helton, University of Oklahoma, University of Oklahoma, USA

Keywords: Ethical behavior, Integrity, Organizational influences, Scientific integrity, Situational influences

Our image of the working scientist remains inherently romantic (1). We envision an individual,
working alone, pursuing knowledge in an area solely for its intrinsic interest. As attractive as the
image may be, it has little to do with the realities of current work in the sciences (2, 3, 4). Scientists
work in a distinctly social setting, conducting their work in both collaboration and competition with
others (5, 6). This work, moreover, occurs in organizational settings, including business, government
and academia. Thus, the pressures that face people working in any organization – pressures of time,
conformity, resources, and production – also confront scientists.
Although one might argue that scientists, by virtue of their work, are granted more autonomy and
are carefully buffered from the more “ugly” demands of organizational life, the conditions currently
confronting most scientific endeavors are such that we can expect organizational pressures to become
a progressively more important influence on scientific work. The emerging forces of the new
economy, where innovation is the true competitive edge, move scientists from the periphery of the
business world to the heart of the industrial enterprise (7). Academia, moreover, under the financial
pressures imposed by funding cutbacks, has placed a new emphasis on responding to the needs of the
business community (8). Finally, academia has begun a slow process, for good or ill, of learning how
to manage itself differently, and manage itself like a business.
Given these pressures, there is a need to understand how organizational variables influence
scientific integrity. Unfortunately, systematic studies of scientific integrity are virtually nonexistent.
However, a number of scholars have sought to understand the variables that influence integrity in
organizational settings as a general phenomenon. Accordingly, our intent in the present study is to
examine prior studies of integrity with respect to their implications for understanding organizational
influences on scientific integrity. We will begin by considering the findings obtained in one line of
research concerned with the individual and situational factors that influence integrity in
organizational settings. Subsequently, we will examine the kind of organizationally-based situational
variables that might influence scientific integrity using a multi-level perspective that considers
situational variables operating at the individual, group, and organizational levels of analysis (9).

Studies of Integrity
Psychological studies of integrity have typically employed one of two broad approaches (10). The
first approach holds that integrity, or the lack thereof, is primarily a function of certain characteristics
of the situation in which people find themselves. Thus, studies along these lines examine the
Corresponding author: Michael D. Mumford, Department of Psychology, University of Oklahoma, 455 W. Lindsey, #705
Dale Hall Tower, Norman, OK 73019, 405-325-5583 (voice), 405-325-4737 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
opportunities provided for dishonest behavior to the emergence of object beliefs, or the view
(11), the reinforcements and punishments that others can be used as tools for personal gain
associated with unethical acts (12), perceptions (14, 22). In harming others, unless such effects
of procedural justice (13), and stress and are inhibited by self-regulation, people are likely
authority norms (14). The second approach to acquire negative images of others and their
holds that a lack of integrity is primarily a relationships with others. Thus, object beliefs,
function of certain characteristics of the along with fear, may lead to the emergence of
individual. Scholars applying this second negative life themes. Negative life themes, along
approach have sought to develop global measures with object beliefs, power motives, self-
of integrity (15, 16), and identify certain unique regulation and outcome uncertainty reflect beliefs
characteristics of people that are associated with and motives held to exert direct effects on
a lack of integrity (17, 18). people’s willingness to engage in destructive
unethical acts. Figure 1 provides a summary of
Individual Variables the key structural relationships specified in this
In one series of studies along these lines, model.
Mumford and his colleagues (19-21) sought to In an initial test of the plausibility of this
develop a general model of the individual model, O’Connor, Mumford, Clifton, Gessner,
characteristics likely to promote destructive or and Connelly obtained biographies for 82 notable
unethical acts. To identify the characteristics of historic leaders (21). They content-coded the
individuals related to the propensity for unethical “rise to power” chapters included in each
acts, Mumford and his colleagues reviewed biography for leaders’ expression of behaviors
relevant studies in the clinical (22-24), indicative of the seven characteristics included in
management ethics (12, 18, 25), social- this model (e.g., object beliefs, narcissism, etc.),
personality (26-28), and criminology (29-31) and obtained indices of the harm done to society
disciplines. This review resulted in the by leaders’ policies. In a subsequent causal
identification of seven individual characteristics modeling effort, not only was support obtained
that might plausibly be related to socially for the ability of these variables to predict harm
destructive unethical behavior: 1) narcissism, done by leaders’ policies, it was found that the a
2) fear, 3) outcome uncertainty, 4) power priori structural model presented in Figure 1
motives, 5) object beliefs, 6) negative life provided adequate fit to the observed data. The
themes, and 7) lack of self-regulation. resulting model is shown in Figure 2.
These differential characteristics were held to In the second set of investigations, Mumford,
operate as a dynamic syndrome in shaping Connelly, Helton, Mowry, and Osburn sought to
unethical acts. It was held that narcissism, or determine whether the variables included in this
extreme self-absorption and overevaluation of the model could account for scores on standard
self leads to a motivated defense of a weak self- measures of integrity (34). Here 292 subjects
system (22, 32). This perception of threat, in were asked to complete two overt measures of
turn, induces outcome uncertainty and activates integrity, the Reid Report (35) and the London
power motives as a defensive strategy. Fear, or House PSI or Personnel Selection Inventory (36).
anxiety, is also held to lead to
perceptions of threat, thereby +

leading to outcome uncertainty

Self -Regulation
(33). When people are uncertain
about their capacity to attain -
-
desired outcomes, self-protective +
Negative Life Themes Object Beliefs
tendencies will activate power +
motives, although the activation of - + +
Lack of
Integrity
power motives may be somewhat -
inhibited by the tendency of Fear Power Motives +

fearful individuals to withdraw. +

+
+
Once activated, power motives +

induce a tendency to harm or Narcissism + Outcome Uncertainty

exploit others which, with the
resulting desensitization, may lead Figure 1. General structural model for individual influences on integrity.

74
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
.35 to complete the background data
scales measuring the beliefs and
Self-Regulation
motives relevant to integrity (e.g.,
.02
- .10 object beliefs, power motives,
.97 etc.). Additionally, manipulations
Negative Life Themes Object Beliefs
.06 were made in the conditions of
Leader
- .04 .97 Destructiveness task performance, specifically
.02 authority norms, psychological
Fear Power Motives .24
distance, and feelings of self-
.83
.55
.81 .26
efficacy. It was found that MBA
.13
students who expressed individual
Narcissism .40 Outcome Uncertainty characteristics held to influence
the occurrence of unethical acts
Figures 2. Test of structural model for individual influences with respect to would take unethical actions when
leader destructiveness.
feelings of self-efficacy were low.
Both these measures examine theft, dishonesty, However, they would not
and punitive attitudes as direct markers of necessarily make unethical decisions unless they
integrity. In addition, 400 subjects were asked to had reason to believe that the actions taken
complete two commonly used personality based would be supported by people in authority. Thus,
measures of integrity (37) – the Socialization and it appears that situational variables might
Delinquency scales of the California influence ethical decisions potentially interacting
Psychological Inventory (CPI). Here background with individual predispositions in conditioning
data scales were developed to measure each of the occurrence of unethical behavior or,
the characteristics included in this model using alternatively, by creating unique effects on
the procedures suggested by Mumford, Costanza, unethical behavior.
Connelly, and Johnson (38). Again, it was found
that the structure of the a priori model was Situational Variables
confirmed. However, here it was found that In fact, beginning with the work of Hartshorne
although scores of these differential variables and May (11), many scholars have argued that
yielded effective prediction of integrity test situational variables might exert strong effects on
scores (r = .32), the obtained prediction was not unethical behavior. In an initial investigation
of overwhelming power. Figure 3 illustrates the intended to identify the kind of situational
nature of the results obtained in this study, while variables that might influence the occurrence of
Table 1 describes the items used to measure these unethical acts, Gessner, O’Connor, Mumford,
variables. Clifton, and Smith developed a set of life history
A potential explanation for the limited, albeit items intended to capture exposure to situations
significant, impact of these variables on integrity likely to influence development, or expression of,
test scores may be found in a study conducted by the various individual characteristics held to
Mumford, Gessner, Connelly,
O’Connor, and Clifton (20). In .07

this study, 152 Masters of Self-Regulation

Business Administration (MBA)
students were asked to work on an - .19
.02
in-basket exercise which presented .37
Negative Life Themes Object Beliefs
32 decisions that might be made .05
by regional sales managers. On - .15 .27 .11 Dishonesty
half of the items included in this - .24 .10
in-basket exercise, the MBA Fear
.46
Power Motives

students were presented with .51

.52
ethical decisions where the actions .43
selected might result in harm to Narcissism .30 Outcome Uncertainty
others or harm to the organization.
Prior to starting work on this Figures 3. Test of structural model for individual influences with respect to
task, the MBA students were asked integrity.
75
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Individual Scales Example Items

Object Beliefs Surprised by how much people invest in friendships; did not do favors for people who
could not return them; told white lies to get own way; viewed dealing with people as a
game; has not gotten emotionally involved when dealing with people.
Power Motives Frustrated when could not convince friends to adopt one’s view; was important to be
on the winning side; was willing to make a scene to get compliance from others;
enjoyed making others do things; liked to have the last word.
Negative Life Themes Enjoyed parties where people were really out of control; was not upset by media
violence; spending time with family was not important; has not reflected upon one’s
purpose in life as much as others.
Outcome Uncertainty Often planned for things that never happened; wished things would slow down or
remain the same; worried about the future; annoyed by people who claimed something
was a sure thing; wished there were more guarantees in life.
Fear Friends thought they worried too much; often agonized over decisions; often woke up
at night for no apparent reason; was bothered by things that could go wrong when
things were going well; had difficulty making decisions about the f uture.
Narcissism Tried to make self look good; was important to receive praise from others; spend a lot
of time worrying about appearance; did not talk about things not of interest to them;
did not spend time with others whose opinions were different.
Lack Of Self-Regulation Not hard on one’s self; rarely said the right thing at the right time; not important to
identify own limitations; took long to fit in with an unfamiliar crowd; did not express
opinions according to the situation at hand.

Table 1: Examples of Items Included in the Individual Scales

influence unethical behavior (e.g., object beliefs, individual characteristics held to influence
outcome uncertainty, etc.) (39). A subsequent unethical behavior (e.g., negative life themes,
factoring of these items after they had been object beliefs, etc.) yielding bivariate correlations
administered to 285 undergraduates, lead to the in the .40s. The second major finding indicated,
identification of seven situational factors: however, that the situational variables were
1) alienation, 2) non-supportive family, 3) nega- strongly related to integrity test scores producing
tive role models, 4) life stressors, 5) competitive relationships in the mid-.20s to low-.50s. Of
pressure, 6) exposure to negative peer groups, these variables, exposure to negative peer groups,
and 7) financial need. Table 2 illustrates the alienation, and financial need appeared to
nature of the items used to measure these produce the strongest relationships across the
variables. four measures of integrity. The third major
To examine the impact of these variables on finding to emerge in these analyses indicated that
integrity, Mumford, Connelly, Helton, Mowry, the situational variables yielded better prediction
and Osburn, administered the life history items of scores on the four integrity tests than the
measuring exposure to these situational factors to individual variables while yielding significant
the 292 subjects asked to complete the two overt gains in prediction when added to the individual
integrity tests, the Reid Report and the PSI, and variables. The results obtained in this third
the 400 subjects asked to complete the two analysis are summarized in Figure 4 which
personality-based tests, the CPI socialization and indicates that the situational variables accounted
delinquency scales (34). In this study, scores on for far more variance in integrity test scores than
the overt and personality based measures of the individual variables.
integrity were both correlated with, and regressed Although these findings underscore the
on, the seven situational scales. fundamental importance of understanding
The first major finding to emerge from these situational influences in attempts to understand
analyses was that the situational scales were and control unethical acts. These findings leave
correlated with scores on the measures of two crucial questions unanswered. First, they do
76
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity

Situational Scales Example Items

Alienation Had worked in a setting where they saw discrimination; had superiors who were
condescending; worked with people and withheld information; had belonged to
organizations in legal trouble; lost something because others took advantage of status or
position; often worked in situations where they could not keep up with demand.
Non-Supportive Family Parents were not consistent in praise of punishment; parents did not explain why they
punished; parents and teachers did not praise work; did not have input into important
family decisions; parents and siblings did not help with schoolwork.
Negative Role Models Parents broke promises; parents openly criticized others; often witnessed violent
arguments among adults in household; parents gave harsh punishments; parents lost
temper for no apparent reason; family had different standards than other families.
Life Stressors Unable to go to school due to health; had to cope with large unexpected expenses;
teachers made unrealistic work demands; had serious illness; schoolwork effected by
problems of family members; was in situations where they could not keep up with work.
Competitive Pressure Often experienced competition among coworkers; concerned about finding a good job
after graduation; frequently sought recognition for work; had to be competitive to get
ahead at work or school; selected people for membership in clubs; was involved in team
projects.
Negative Peer Group Friends had a cynical attitude towards society; high school and college friends had
trouble with law; friends and family were heavy users of drug and alcohol; observed
people breaking rules while growing up; saw people taken advantage of; witnessed
verbal/physical violence.
Financial Need Many families in neighborhood they grew up in received some type of public assistance;
lost mother or father; regular schedule was not emphasized in family; members of family
had been in trouble with law; people could take things away from them because of family
position.

Table 2: Examples of Items Included in the Situational Scales

not tell us exactly how unethical acts are deadlines, the need to acquire resources, and
influenced by situational variables. For example, uncertainty about project outcomes (40). When
situational variables might constrain unethical these occupational demands are combined with
behavior, interact with individual variables or, the intense focus characteristic of those engaged
alternatively, compel unethical behavior in their in scientific work (41), it seems plausible to
own right. Second, these findings do not tell us argue that stress represents an endemic feature of
about the specific kinds of situational variables life in the sciences. Although, up to a point,
that act to influence unethical behavior in the stress may contribute to productivity, high levels
kind of organizational settings in which scientists of stress may not only prove debilitating, but,
are likely to work. Accordingly, in the following more centrally, may contribute to incidents of
sections, we will examine the specific kinds of unethical conduct through two distinct
situational variables operating at the individual, mechanisms (42). First, high levels of stress may
group, and organizational levels that might lead people to take more risky actions than they
influence scientific integrity. might under other conditions due to the negative
effects of stress on self-regulation (27). Second,
Individual Level stress reduces the cognitive resources available
Of the situational variables found to be related to for reasoning and analytical problem solving
integrity, stress seems to be the variable most (43). This loss in cognitive capacity is
likely to be linked to integrity in research work. noteworthy because effective moral reasoning
Scientific work is known to be demanding and inhibits the occurrence of unethical acts (18, 44,
stressful resulting from multiple commitments, 45). These observations, in turn, lead to our first
77
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Personality Based Tests Overt Tests

CPI CPI PSI Reid PSI Reid
Socialization Delinquency Honesty Honesty Theft Theft
INDIVIDUAL SCALES
Multiple Correlations .42 .38 .36 .27 .25 .30
Cross Validated Multiple Correlation .36 .31 .29 .20 .07 .17
SITUATIONAL SCALES
Multiple Correlation .62 .58 .57 .43 .35 .28
Cross-Validated Multiple Correlation .49 .51 .40 .38 .26 .12
SITUATIONAL SCALES ADDED TO
INDIVIDUAL SCALES
Multiple Correlation .67 .61 .61 .47 .41 .40
Cross-Validated Multiple Correlation .62 .50 .58 .38 .27 .17
Change in R Square .26** .23** .24** .17** .11** .07*

Figure 4: Comparison of Individual and Situational Variables with Respect to the Prediction of Integrity Test Scores.
*P < .05 ** P < .01
two propositions. Experienced, competent workers, however, may
• Proposition One: Incidents of unethical behavior also feel less need to take shortcuts. Regardless
will be more frequent when individuals experi- of the explanation used to account for these
ence stress and overload. effects, however, it is clear that organizations
• Proposition Two: Attempts by organizations to may take a number of steps to build competence
reduce stress by minimizing time pressure, and expertise through educational and mentoring
managing overload, clarifying goals, and programs, careful selection of employees, and
providing requisite resources will reduce inci- providing people with time to pursue continuing
dents of unethical behavior. education projects (2).
Actions taken to reduce work demands, of Competence and expertise, of course, also
course, are not the only steps that might be taken allow people to induce effective control over
to reduce stress and unethical behavior in their work environment. Given the impact of
organizational settings. Both stress and stress, outcome uncertainty, and fear on unethical
uncertainty about outcomes are influenced by acts, one would expect that control beliefs would
people’s feelings of competence and their ability be related to unethical behavior in organizational
to exert positive, effective control over their settings. In fact, studies by Hegarty and Sims
work environment. In keeping with this (12), Trevino and Youngblood (18), and Reiss
observation, Weeks, Moore, McKenney, and and Mitra (51) all indicate that people who have
Longnecker administered vignettes calling for a strong internal locus of control are less likely to
ethical decisions to managers with greater and engage in unethical acts than people who believe
lesser experience (46). They found that their actions are controlled by external forces.
experienced managers were more likely than What is important to recognize here, however, is
their less experienced counterparts to make that organizations can build feelings of control by
ethical decisions. Other studies by Arlow and assigning people to tasks commensurate with
Uhlrich (47), Chonko and Hunt (48), Kidwell, their capabilities, allowing input to critical
Stevens, and Bethke (49), and Teal and Carroll decisions, and buffering people from
(50) also indicate that more experienced uncontrollable events. Taken as a whole, these
successful workers, workers with greater observations imply the following three
expertise, are less likely to engage in unethical propositions.
activities or make unethical decisions. As noted • Proposition Three: Less skilled or less experi-
above, one potential explanation for these enced scientists will be more likely to engage in
findings is the ability of experienced, competent unethical acts and will be more sensitive to
workers to handle stress and uncertainty. organizational pressures that promote unethical
acts.
78
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
• Proposition Four: Organizational actions the potential contribution of its’ worth to society
intended to develop expertise and maximize as a whole (55, 56). As Bowie points out, this
feelings of competence will inhibit unethical acts. intrinsic motivation buffers individuals from
• Proposition Five: Organizational actions in- situational pressures likely to promote unethical
tended to maximize people’s control of their acts (57) . He notes, furthermore, that a variety of
environment will inhibit unethical acts. organizational policies might influence alienation
As important as competence and control may and intrinsic motivation including explicit
be to the management of stress and the recognition of social contributions as well as
minimization of unethical behavior, some contributions to the “bottom line”, allowing
consideration should be given to family and individuals to pursue personally interesting work,
social relationships. Family and social and maximizing autonomy in decision-making.
relationships, specifically supportive These observations suggest the following
relationships, help people cope with stress while proposition.
the implied commitment to others embedded in • Proposition Eight: Attempts by the organization
these relationships promotes a prosocial outlook. to recognize and reward social contributions and
Accordingly, Mumford, Connelly, Helton, allow individuals to pursue their unique interests
Mowry, and Osburn (34) found that exposure to a will reduce incidents of scientific misconduct.
non-supportive family environment was related Eisenberger and Cammeron, however,
to a lack of integrity. Unfortunately scientists, in remind us that creative work, including scientific
part due to their introversion (52) and, in part due work, is not simply a matter of intrinsic
to their work commitments (53), appear to have motivation (58). People’s work as scientists is
some difficulty in establishing viable family and also motivated by extrinsic factors such as pay,
social relationships. By the same token, recognition, and status. At first glance, it might
however, scientists do appear to establish viable, seem plausible to argue that extrinsic rewards
long-term collaborative relationships and create lead to unethical behavior. However, the
social connections through their network of relationship between the pursuit of extrinsic
enterprise (5, 54). These observations, in turn, rewards and unethical behavior appears
suggest that incidents of unethical behavior will somewhat more complex with the pursuit of
occur less frequently among scientists who have extrinsic rewards contributing to unethical acts
a rich extensive network of supportive only when people expect that the unethical
professional colleagues. Moreover, by co- behavior will be rewarded, the unethical act will
locating scholars with similar interests, not be detected, and the act, if detected, will not
encouraging collaborative work, recognizing the be sanctioned by the organization (12, 18, 59).
value of multiple-authored publications, and One implication of this expectancy model is that
providing time for collegial interactions, high performers will sometimes engage in
organizations can reduce incidents of scientific unethical acts because they believe they are less
misconduct. Thus: likely to be sanctioned by the organization (60,
• Proposition Six: Individuals lacking collabora- 61)–potentially resulting in a culture that seems
tive networks will be more likely to be involved to condone such acts. Another implication of this
in incidents of scientific misconduct. expectancy model is that ethical behavior will
• Proposition Seven: Organizational actions decrease when extrinsic rewards such as pay and
intended to facilitate and recognize the value of promotions are based on immediate short-term
collaborative activities will minimize incidents of production demands rather than long-term
scientific misconduct. contributions to others (62).
Our foregoing observations with regard to In considering the impact of production
collaboration point to another factor likely to be demands, however, it is necessary to bear in mind
involved in incidents of scientific misconduct – a unique characteristic of scientific work.
alienation. Alienation among scientists is not a Scientists’ rewards are often explicitly tied to
strictly social phenomenon. Alienation from the production such as journal publications, patents,
work, and the work’s potential contributions to and fielding new software (63, 64). By expressly
society, appear particularly significant with tying extrinsic rewards to production counts,
regard to scientific misconduct because scientific however, one can expect that misconduct will
work is often motivated by intrinsic interest in increase whenever ambitious, extrinsically
the work for its own sake and an abiding belief in motivated individuals, individuals motivated by
79
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
financial needs, status concerns, and recognition, leadership styles, specifically consideration and
encounter significant reverses in the production initiating structure. They found that the leaders’
process. Thus, organizations might minimize emphasis on initiating structure contributed to
misconduct by rewarding progress towards goals ethical decision-making, presumably because the
as well as production output, recognizing initiation of structure led group members to focus
alternative indices of performance such as impact on task accomplishment rather than personal
and innovation, and providing a minimal degree concerns. In another study along these lines,
of security and visibility for all group members Zabid and Alasgoff found that the behavior of
based on their unique strengths.(65) Taken as a people’s immediate superior exerted stronger
whole, our preceding observations about effects on the occurrence of unethical acts than
extrinsic motivation suggest the following four other putative organizational influences such as
propositions. climate and codes of conduct (68).
• Proposition Nine: Organizational reward systems Leaders appear to influence ethical behavior
that stress long-term innovation and impact will through a variety of different mechanisms, some
tend to minimize incidents of unethical behavior. of which may inhibit unethical acts and some of
• Proposition Ten: Organizational reward systems which may promote such acts. Sims, in a study
that recognize progress as well as output will of leadership in financial services firms,
tend to minimize incidents of unethical behavior. identified four ways leadership behavior
• Proposition Eleven: Scientific misconduct will contributes to or promotes integrity (69). He
occur more frequently when extrinsic rewards are argues that leaders promote ethical behavior by
based on production and people are treated a) focusing the attention of people on ethical
harshly for production setbacks. issues, b) responding to crises based on ethical,
• Proposition Twelve: Scientific misconduct will productive concerns rather than self-protection,
occur less frequently in organizations where all c) allocating rewards based on long-term
incidents of misconduct are treated similarly, contributions rather aggressive self-promotion,
regardless of the past performance of the people and d) applying sanctions for incidents of
involved. unethical behavior. Along similar lines, Minkes,
Small, and Chatterjee have argued that leaders’
Groups articulation and communication of personal,
The Mumford, Connelly, Helton, Mowry, and ethical, and moral values will promote integrity
Osburn study not only points to the influence of on the part of group members (70). Contrawise,
individual level situational influences on it appears that leaders who articulate poor values
integrity, such as stress, relational support, or exhibit self-serving, narcissistic behavior
alienation, and financial need, it also underscores implicitly encourage unethical behavior on the
the importance of certain group level influences part of subordinates (71, 72). Vredenburgh and
(34). In this study, three variables operating at Brender point out, moreover, that leaders who
the group level, role models, exposure to consistently abuse power through arbitrary
negative peer groups, and competitive pressure, actions, a focus on personal control, and
were found to influence integrity. Again, all inequitable decisions, induce stress, fear, and
three of these situation variables appear to outcome uncertainty while activating the power
represent important influences on integrity in motive linked to unethical acts (73).
organizational settings. Although it seems clear that leaders have an
In organizations, role modeling is commonly impact on ethical behavior in general, the
subsumed under this broader area of leadership question remains as to whether leaders have a
(66), and there is, in fact, reason to believe that similar impact on the ethical behavior of
the behavior of people assigned to formal scientists. One might argue that, due to their
organizational leadership roles will influence the greater autonomy and specialized professional
manifest integrity of their “followers”. In one expertise, scientists are less susceptible to leader
study along these lines, Schminke and Wells had influence (66, 74). Although this argument
81 business students participate in a four-month seems plausible, the available evidence indicates
long strategic planning simulation (67). During that leaders exert notable effects on people’s
the course of this simulation, measures of ethical behavior in research settings (75). A case in
decision-making were obtained along with point may be found in Hounshell’s analysis of
measures of group process variables and research on synthetic fabrics in Dupont’s Pioneer
80
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
research laboratories where the vision defined by acts. Scientists have been found to be highly
founders in the 1920s continued to shape the competitive evidencing not just competitive
laboratories’ research programs well into the intensity but some degree of hostility and
1990s (76). Nonetheless, the autonomy and arrogance (79)–all dispositional factors likely to
expertise of scientists suggest that leader make scientists particularly susceptible to the
influences on ethical issues will be less evident in negative effects of competitive pressure.
day-to-day direction and more evident in the Competitive pressure, however, may not always
leaders’ be destructive provided it is managed effectively
a) definition of a coherent constructive by the organization (80). More specifically,
research vision, b) focus on production as when competition is accompanied by respect for
opposed to status relationships, and c) articula- competitors, people feel that they have sufficient
tion of ethical values in interactions with staff. technical competence to compete effectively, and
When these observations are considered with competition is viewed as a depersonalized,
respect to the findings sketched out above, they professional challenge, then competition may
suggest the following three propositions: contribute to performance and ethical behavior
• Proposition Thirteen: Scientific misconduct will (81, 82). These observations, in turn, suggest the
be less common in groups where leaders have the following three propositions.
expertise needed to define a coherent vision for • Proposition Sixteen: Unethical acts are more
the work. likely to be observed when ambitious, highly
• Proposition Fourteen: Scientific misconduct will competitive people are placed in competitive
be less common in groups where the leader settings where they lack requisite skills.
actively articulates ethical values, potential social • Proposition Seventeen: Organizations that take
contributions of the work, and enhancement of actions to reduce personalized competitive
the work rather than career status. pressure by evaluating performance on an
• Proposition Fifteen: Scientific misconduct will absolute rather than relative basis or by encourag-
be less common in groups where the leader ing collaborative work among potential competi-
focuses on effective direction of production tors are less likely to experience incidents of
activities rather than personal professional unethical behavior.
recognition, maintenance of control, or social • Proposition Eighteen: Unethical behavior is less
acceptance. likely to occur when leaders, or organizational
Leadership, of course, is not the only group practices, encourage people to analyze and
level variable that might influence integrity in identify the merits in competitors’ work.
organizational settings. For example, Mumford, Personalized competition within-groups, of
Connelly, Helton, Mowry, and Osburn found that course, may result in conflict and a lack of
competitive pressure was related to a lack of cohesiveness. In this regard, the Schminke and
integrity (34). The effects of competition on Wells study cited earlier is noteworthy. In
ethical behavior, however, appear to be quite addition to examining leadership styles and their
complex in organizational settings. One way influence on ethical decision-making, they also
competition appears to influence ethical behavior examined the effects of group cohesiveness (67).
may be found in the tendency of people to Here it was found that cohesiveness influenced
discount the relevance of moral considerations to ethical decision-making both directly with more
decision-making in competitive situations (77). cohesive groups making more ethical decisions
Another way competition influences ethical and indirectly with cohesive groups evidencing
behavior is that negative perceptions of higher performance which, in turn, led to more
competitors’ intentions provide a justification of ethical decision-making. These findings suggest
unethical acts (78). Still another way that actions taken to induce cohesiveness through
competition influences ethical behavior is by development and articulation of a shared,
inducing feelings of stress and uncertainty (39). common vision, use of group as well as
These varied mechanisms by which individual rewards, integration of members work
competition influences ethical behavior are all activities, and encouragement of within-group
clearly applicable to scientists. In the case of collaborative efforts will all contribute to ethical
scientists, however, it is quite possible that these behavior. Thus, the following three propositions
negative aspects of competition represent seem indicated.
particularly important influences on unethical
81
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

• Proposition Nineteen: Unethical acts are more ethical behavior among group members will
likely to occur in non-cohesive conflict-laden contribute to integrity. As might be expected, the
groups. bulk of the available evidence does indicate that
• Proposition Twenty: Cohesiveness within a ethical norms within a group lead to ethical
group will reduce scientific misconduct both by behavior. For example, studies by Barnett (86),
enhancing performance and minimizing the Kawathatzopoulos (87), Verbke, Ouwerkerk, and
negative effects of within-group competition. Peelen (88), and Weaver and Farrell (89) indicate
• Proposition Twenty-One: Organizational actions that when groups communicate expectations for
that lead to higher cohesiveness, such as develop- ethical behavior, and sanction violations by
ment of a shared vision on the allocation of group members, ethical decision-making
group, as well as individual, rewards, will reduce improves and unethical acts become less
incidents of scientific misconduct. frequent. In this regard, however, it is important
Although it appears that cohesiveness may to bear in mind a point made by Fritz, Arnett, and
contribute to integrity, a cautionary note seems in Conkel (90), Grimalda (91), and Schokkaert and
order. Many prior studies of groups, including Sweeney (92). More specifically, the effects of
destructive behavior on the part of groups, group norms on ethical behavior will vary with
indicate that conformity pressures can induce people’s commitment to the group. Accordingly,
destructive, unethical behavior when the primary the following three propositions seem indicated.
concern is maintenance of harmonious group • Proposition Twenty-Four: Ethical behavior will
relations and the goals being pursued by the be more common in groups that have, and
group are likely to result in destructive, unethical actively apply, positive normative standards in
behavior (24, 83). Hence: group decision-making and the application of
• Proposition Twenty-Two: When high levels of sanctions.
cohesiveness prohibit questioning of group • Proposition Twenty-Five: The effects of ethical
actions, cohesiveness may be related to unethical norms on integrity depend on building feelings of
acts. commitment to the group, the organization, or the
As implied by our foregoing proposition, profession.
exposure to the behaviors of, and expectations • Proposition Twenty-Six: the creation and
imposed by, other group members may influence articulation of normative ethical standards by
ethical behavior in organizational settings (34). leaders on professional organizations will prove
Exposure to peer groups is commonly held to less effective when groups are experiencing rapid
influence integrity through the models for change and commitment is low.
appropriate behavior provided by other group
members and the normative expectations Organizations
imposed on people by other members of the The Mumford, Connelly, Helton, Mowry, and
group (39, 84) . Accordingly, Murphy has argued Osburn study focused primarily on situational
that anomie, or normlessness, will engender factors operating at the individual or group level
unethical behavior because group members lack (34). As a result, this study does not directly
models for appropriate behavior and sanctions address the various organizational level variables
are not imposed for unethical acts (10). In that might be related to integrity. Nonetheless,
keeping with argument, Leede, Nijhof, & the nature of the individual and group based
Fisscher, note that when groups are experiencing situational influences on integrity do suggest that
conditions of rapid change the resulting certain organizational level variables will also
breakdown in extant normative structures may influence integrity. One set of organizational
lead to an increase in the frequency of unethical level influences suggested by our foregoing
acts (85). Thus, observations is the organization’s operating
• Proposition Twenty-Three: When groups are environment – specifically three features of the
experiencing rapid changes in personnel, technol- organization’s operating environment turbulence,
ogy, or productions processes, incidents of munificence, and interdependence.
unethical behavior will increase. Environmental turbulence refers to rapid
The notion that normlessness will contribute changes in technology, business processes,
to the occurrence of unethical acts also implies product markets, and competitors (93). Of
that the presence of normative expectations for course, turbulence will lead to normlessness as
well as uncertainty about the requirements for
82
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
effective performance, both conditions that can partners, or government agencies. As might be
be expected to promote unethical acts. expected, high interdependence appears to
Accordingly, Morris, Marks, Allen, and Perry promote ethical behavior (99, 100, 101).
found that ethical values were less evident among Although it is unclear exactly what mechanisms
people working for organizations operating in a shape the influence of interdependence on ethical
turbulent environment (94). Along similar lines, behavior the following proposition does seem
Rossouw has argued that the turbulence induced indicated:
by social disruption can lead to unethical acts on • Proposition Thirty: Unethical behavior occurs
the part of organizations (95). Among scientists, less frequently in organizations where perfor-
however, it seems likely that turbulence will mance depends on the support, or goodwill, of
exert larger effects when its impact is evident in other entities.
their immediate technical environment or in The organization’s operating environment is,
employment practices. These observations, in of course, one influence on the structure of the
turn, lead to the following two propositions. organization. Structure, or the manifest division
• Proposition Twenty-Seven: As turbulence of labor in an organization, has not commonly
increases in the organization’s operating environ- been studied as an influence on integrity.
ment the frequency of unethical acts will in- However, the available evidence indicates that
crease. unethical acts are less likely to occur in small
• Proposition Twenty-Eight: Scientific misconduct organizations (102, 103) and in organizations
will increase in periods of rapid change in where roles and responsibilities are clearly
technological paradigms and employment defined (85, 104). One explanation for this
practices. pattern of findings may be found in diffusion of
In contrast to turbulence, munificence refers responsibility and its derivative effects or
to the availability of resources and the low alienation. In keeping with this alienation and
degree of competitive pressure evident in the diffusion of responsibility notion, Dooley and
organizations’ operating environment. In fact, Fryxell found that diversification was related to
the available evidence indicates that munificence corporate pollution levels (105). These
is related to ethical conduct in organizational observations imply the following proposition:
settings. For example, Verschoor (96), in a study • Proposition Thirty-One: As organizational
of Fortune 500 companies, found that ethical structures become more complex, and roles and
conduct with regard to organizational role accountability are less clearly defined for
shareholders increased with financial individuals, unethical acts will become more
performance while Judge (97), in a study of frequent.
hospitals, found that scarcity of financial While structure refers to the organization of
resources was negatively related to social the work, climate refers to people’s perceptions
contributions. In still another study along these of social interactional expectations with their
lines, Zarkada-Fraser found that collusion in work environment (106). Relative to structure,
government project bids was related to project climate has received substantially more attention
desirability and competition (98). Among as a potential influence on ethical behavior in
scientists, where resources are critical to organizational settings. In one study along these
conducting requisite research work, non- lines, Sims and Keon administered five business
munificent environments may encourage scenarios calling for an ethical decision to 245
unethical acts as a way of insuring resource business students who were also asked to
availability. Thus, complete a survey describing the company for
• Proposition Twenty-Nine: As the munificence of which they were currently working (107). It was
the organizations operating environment de- found that perceptions of their work environment
creases, unethical behavior and incidents of were related to ethical decision-making. Similar
scientific misconduct will increase. findings have been obtained by Baumhart (59).
A third, and final, environmental variable Although there is reason to believe that
commonly linked to ethical behavior in organizational climate influences ethical
organizational settings is interdependence, or the behavior, more debate surrounds the nature of the
extent to which organizational success depends specific climate dimensions involved. Agarwal
on maintaining viable relationships with other and Malloy identify five climate dimensions
organizations including suppliers, alliance related to ethical behavior:
83
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
1) individual caring 2) social caring, 3) inde- On the other hand, however, there is no
pendence, 4) Machiavellianism, and 5) law and assurance that professional ethical codes will be
code (108). Vidaver-Cohen proposes a different adopted by organizations in their day-to-day
model of ethical climate which stresses the practices. This point is nicely illustrated in a
importance of 1) social responsibility, 2) social study by Etheredge who examined attitudes
support, 3) avoiding harm of others, 4) task toward ethical behavior in business managers and
support, and 5) equity of reward procedures identified two dimensions: a) the importance of
(109). Still another model, one proposed by Key, ethics and social responsibility, and
views climate as a function of: 1) day-to-day b) subordination of ethics and social
reinforcement of ethical conduct, 2) punishment responsibility to organizational effectiveness
of unethical conduct, and 3) management role (115). Thus, organizations in their quest for
modeling (110). Finally, Argadona and Hartman, efficiency and control, may reject professional
Yrle, and Galle argue that trust and perceptions ethical standards that conflict with organizational
of distributive and procedural justice represent needs. When organizations reject these
key organizational climate dimensions professional standards, however, it can be
influencing ethical behavior on organizations expected that the resulting organizational-
(111,112). professional conflict will induce some stress as
While a variety of models of ethical climate people are forced to choose between these
are available, it seems likely that some of these competing expectations. Although a number of
dimensions will prove more important than considerations will influence how this conflict is
others in shaping the ethical behavior of resolved, it appears that investment in the
scientists. Given the hostility and organization, as opposed to the profession, is of
competitiveness characteristic of scientists (79), critical importance (116). Accordingly, the
it seem plausible to argue that climates stressing following three propositions seem indicated.
trust and social support while maintaining • Proposition Thirty-Four: Incidents of scientific
perceptions of procedural and distributive justice misconduct will be less common among indi-
will prove particularly important in minimizing viduals who are more invested in the profession
misconduct (7). The demands of creative work, rather than the organization they are working.
moreover, suggest that climates reinforcing • Proposition Thirty-Five: Incidents of scientific
autonomy, openness, and minimization of misconduct will be less common in organizations
premature criticism will also prove useful in that rely on their professional technical reputation
enhancing ethical behavior (75, 113). Thus, the for market advantage and view organizational
following two propositions seem indicated. needs as consistent with professional ethical
• Proposition Thirty-Two: Organizational climates codes.
that promote perceptions of trust and fairness will • Proposition Thirty-Six: Professional ethical
minimize incidents of scientific misconduct. codes will prove most effective in reducing
• Proposition Thirty-Three: Organizational scientific misconduct when codes are actively
climates that are open and not overly critical of supported by the organization.
new ideas will minimize incidents of scientific
misconduct. Conclusions and Directions
The climate literature, however, also Figure 5 summarizes the various propositions we
underscores the importance of day-to-day have proposed with respect to the situational
reinforcement on ethical conduct. In the case of variables influencing ethical behavior at the
scientists, the importance of ethical standards individual, group, and organizational levels. In
implies that professional codes, as well as their reviewing these propositions, however, an
acceptance and embodiment by the organization, important caveat seems in order. More
will also influence incidents of scientific specifically, although all of the propositions were
misconduct. In fact, studies by Weaver and formulated based on a review of the
Farrell (89) of American Marketing Association organizational literature as it relates to the
members, and Gotterbarn (114) of software situational variables influencing integrity. Few, if
engineers, indicate that professional codes are any, studies have directly examined the influence
viewed as important influences on ethical of organizational, situational variables on
behavior in the sciences and may lead to research integrity. Thus, these propositions
improvements in ethical decision-making. should not be viewed as well established
84
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity

Individual Level Group Level Organizational Level

1) Incidents of unethical behavior will 13) Scientific misconduct will be less common in 27) As turbulence increases in the
be more frequent when individuals groups where leaders have the expertise needed organization’s operating
experience stress and overload to define a coherent vision for the work environment, the frequency of
2) Attempts by organizations to reduce 14) Scientific misconduct will be less common in unethical acts will increase
stress by minimizing time pressure, groups where the leader actively articulates 28) Scientific misconduct will
managing overload, clarifying ethical values, potential social contributions of increase in periods of rapid
goals, and providing requisite the work and enhancement of the work rather change in technological
resources will reduce incidents of than career status paradigms and employment
unethical behavior 15) Scientific misconduct will be less common in practices
3) Less skilled or less experienced groups where the leader focuses on effective 29) As the munificence of the
scientists will be more likely to direction of production activities rather than organization’s operating
engage in unethical acts and will be personal professional recognition, maintenance environment decreases, unethical
more sensitive to organizational of control, or social acceptance behavior and incidents of
pressures that promote unethical 16) Unethical acts are more likely to be observed scientific misconduct will
acts when ambitious, highly competitive people are increase
4) Organizational actions intended to placed in competitive settings where they lack 30) Unethical behavior will occur
develop expertise and maximize requisite skills less frequently in organizations
feelings of competence will inhibit 17) Organizations that take actions to reduce where performance depends on
unethical acts personalized competitive pressure by evaluating the support, or goodwill, of other
5) Organizational actions intended to performance on an absolute rather than relative entities
maximize people’s control of their basis or by encouraging collaborative work 31) As organizational structures
environment will inhibit unethical among potential competitors are less likely to become more complex, and roles
acts experience incidents of unethical behavior and role accountability are less
6) Individuals lacking collaborative 18) Unethical behavior is less likely to occur when clearly defined for individuals’
networks will be more likely to be leaders, or organizational practices, encourage unethical acts will become more
involved in incidents of scientific people to analyze and identify the merits in frequent
misconduct competitors’ work 32) Organizational climates that
7) Organizational actions intended to 19) Unethical acts are more likely to occur in non- promote perceptions of trust and
facilitate and recognize the value of cohesive, conflict-laden groups fairness will minimize incidents
collaborative activities will 20) Cohesiveness within a group will reduce of scientific misconduct
minimize incidents of scientific scientific misconduct both by enhancing 33) Organizational climates that are
misconduct performance and minimizing the negative open and not overly critical of
8) Attempts by organizations to effects of within group competition new ideas will minimize
recognize and reward social 21) Organizational actions that lead to higher incidents of scientific
contributions and allow individuals cohesiveness such as development of a shared misconduct
to pursue their unique interests will vision or the allocation of group as well as 34) Incidents of scientific
reduce incidents of scientific individual rewards will reduce incidents of misconduct will be less common
misconduct scientific misconduct among individuals who are more
9) Organizational reward systems that 22) When high levels of cohesiveness prohibit invested in the profession rather
stress long-term innovation and questioning of group actions, cohesiveness may than the organization for which
impact will tend to minimize be related to unethical acts they are working
incidents of unethical behavior 23) When groups are experiencing rapid changes in 35) Incidents of scientific
10) Organizational rewards that personnel, technology, or production progress, misconduct will be less common
recognize progress as well as output incidents of unethical behavior will increase in organizations that rely on their
will tend to minimize incidents of 24) Ethical behavior will be more common in professional or technical
unethical behavior groups that have, and actively apply, positive reputation for market advantage
11) Scientific misconduct will occur normative standards in group decision-making and view organizational needs as
more frequently when extrinsic and the application of standards consistent with professional
rewards are based on production 25) The effects of ethical norms on integrity may ethical codes
and people are treated harshly for depend on building feelings of commitment to 36) Professional ethical codes will
setbacks the group, organization or profession prove most effective in reducing
12) Scientific misconduct will occur 26) The creation and articulation of normative scientific misconduct when
less frequently in organizations ethical standards by leaders in professional codes are actively supported by
where all incidents of misconduct organizations will prove less effective when the organization
are treated similarly regardless of groups are experiencing rapid change and
past performance commitment is low
Figure 5. Summary of Propositions at Individual, Group, and Organizational Levels
conclusions but, instead, as a set of hypotheses effort all seem plausible, evidence is not
that might be used to guide further research. available examining the relative importance of
The need for further research along these these various situational variables on scientific
lines becomes even more salient when one takes misconduct and research integrity. For example,
two other considerations into account. First, given the known dispositional characteristics of
although the propositions presented in the present scientists (79), it seems attractive to argue that
85
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
competition, conflict, and a lack of cohesiveness integrity at the group level. These individual and
will have a greater impact on misconduct than group level situational influences, moreover,
the direction provided by a leader. Unfortunately, appear to be associated with a coherent set of
however, evidence allowing us to evaluate the organizational level influences such as turbulence
relative importance of various situational and munificence.
influences within and across three levels of In identifying the situational variables
analysis is, at this juncture, simply not available. operating at the individual, group, and
Second, in formulating these propositions we organizational levels, moreover, it becomes
have examined organizations as a general possible to draw inferences about the conditions
phenomenon drawing heavily from past research under which incidents of misconduct are most
in the “for profit” business arena (18, 107). What likely to be observed and the actions that might
must be recognized here, however, is that be taken by organizations to reduce incidents of
scientists’ work occurs in a variety of settings misconduct. For example, support appears to be
aside from the business arena including related to misconduct with individuals lacking
universities, government agencies, and non-profit collaborative networks and broader social
research institutes. As a result, the unique support being more vulnerable to misconduct.
characteristics of these non-business settings may Organizations, however, by encouraging people
influence the relative importance of the various to collaborate and build a strong network of
situational variables identified in the present professional connections, may do much to
effort. A case in point can be found in our minimize misconduct. Similarly, while
observations about organizational conflicts with competitive pressure apparently plays a notable
professional codes of ethics since such conflicts role in scientific misconduct, such simple
maybe less pronounced outside the business strategies as avoiding person-to-person
setting. Thus, there is a need to assess the comparisons and insuring adequate resources are
generality of these propositions across work available may do much to minimize the
settings. occurrence of misconduct. Hopefully, the
Even bearing these caveats in mind, present effort will serve not only as a framework
however, we believe that the present study does for further research examining the impact of
lead to some noteworthy conclusions about situational variables on scientific misconduct but
research integrity. To begin, we tend to attribute will provide a basis for formulating new policies
incidents of misconduct to characteristics of the that will help insure the integrity of the research
individual. Although the importance of the process. In fact, given the changes occurring in
scientist’s character is not to be underestimated, many scientific fields, there may well in the
the results obtained in the present effort suggest future be an even more pressing need for
that situational variables have a large, perhaps a practical guidelines along these lines as the
larger, impact on integrity than individual rarefied world of science comes into ever closer
variables. Although this argument is by no contact with the manifold demands and pressures
means unique (11), it does suggest that future of the modern organization.
studies of research integrity should give as much
attention to situational and individual influences. Acknowledgements
The present effort, moreover, has served to We would like to thank Shane Connelly, Ted
identify an initial set of situational variables that Gessner, Jennifer O’Connor, and Howard Timm
should be examined in studies of research for their contributions to the present effort. Parts
integrity. The Mumford, Connelly, Helton, of this effort were supported by a series of grants
Mowry, and Osburn study underscores the from the United States Office of Naval Research,
importance of stress, alienation, support, need, Michael D. Mumford, Principal Investigator.
role models, peer groups, and competitive Correspondence concerning this article should be
pressure (34). In this paper we have provided addressed to Dr. Michael D. Mumford,
some evidence that these same situational Department of Psychology, The University of
pressures might also be operating in Oklahoma, Norman, OK 73019, or
organizational settings. For example, stress [email protected] .
appears to be a potentially significant influence
on incidents of misconduct at the individual level
while competitive pressure appears to influence
86
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity

Bibliography 19. Holt RW, Clifton TC, O’Connor JA, Smith TC, Gessner
1. Mumford MD, Baughman WA, Sager CE. Picking the TC, Mumford MD. Influences of destructive
right material: Cognitive processing skills and their personality information on juror decision making. J
role in creative thought. In: Runco MA, ed. Critical Appl Social Psychol. 1997;27:781-99.
Creative Thought. Cresskill, NJ: Hampton; 2000. 20. Mumford MD, Gessner TE, Connelly MS, O’Connor
2. Mumford MD, Gustafson SB. Creativity syndrome: JA, Clifton TC. Leadership and destructive acts:
Integration, application, and innovation. Psychol Bull. Individual and situational influences. Leadership Q.
1985;103:27-43. 1993;4:115-48.
3. Policastro E, Gardner H. From case studies to robust 21. O’Connor JA, Mumford MD, Clifton TC, Gessner TE,
generalizations: An approach to the study of creativity. Connelly MS. Charismatic leaders and destructiveness:
In: Sternberg RJ, ed. Handbook of Creativity. A historiometric study. Leadership Q. 1995;6:529-55.
Cambridge, England: Cambridge University Press; 22. Fromm E. The Anatomy of Human Destructiveness.
1999. New York, NY: Holt, Rhinehart, & Winston; 1973.
4. Weber RJ, Perkins DN. Inventive Minds: Creativity in 23. Gough HG. A sociological theory of psychopathy. Am
Technology. Cambridge, England: Cambridge J Sociology. 1998;53:359-66.
University Press; 1992. 24. Staub E. The Roots of Evil: The Origins of Genocide
5. Abra J. Collaboration in creative work: An initiative for and Other Group Violence. Cambridge, MA:
investigation. Creativity Res J. 1994;7:1-20. Cambridge University Press; 1989.
6. Dunbar K. How scientists really reason: Scientific 25. Robin DP, Reidenbach RE, Forrest PJ. The perceived
reasoning in real world laboratories. In: Sternberg RJ, importance of an ethical issue on the ethical decision
Davidson JE, eds. The Nature of Insight. Cambridge, making of ad managers. J Business Ethics. 1996;35:17-
MA: MIT Press; 1995:365-96. 28.
7. Ruppel CP, Harrington SJ. The relationship of 26. Blass T. Understanding behavior in the Milgram
communication, ethical work climate, and trust to Obedience experiment. The role of personality,
commitment and innovation. J Business Ethics. situations, and interactions. J Pers Soc Psychol.
2000;17:1563-71. 1991;60:348-413.
8. Press E, Washburn J. The Kept University. Atlantic. 27. Leith KP, Baumeister RF. Why do bad moods increase
2000;285:39-55. self-defeating behavior? Emotion, risk-taking, and self-
9. Dansereau F, Yammarino FJ. Introduction and regulation. J Pers Soc Psychol. 1996;71:1250-67.
overview. In: Dansereau F, Yammarino FJ, eds. 28. Oyserman D, Markus HR. Possible selves and
Leadership: the Multiple Level Approaches. Stanford, delinquency. J Pers Soc Psychol. 1990;59:112-25.
CT: JAI Press; 1998:xxv-li. 29. DiLalla LF, Gottesman II. Heterogeneity of causes for
10. Murphy KA. Honesty in the Work Place. Pacific delinquency and criminality: Lifespan perspectives.
Grove, CA: Brooks/Cole; 1993. Dev Psychopathology. 1990;1:339-49.
11. Hartshore H, May MA. Studies in Deceit. New York: 30. Reid JB, Patterson GR. The development of antisocial
MacMillian; 1928. behavior patterns in childhood and adolescence. Eur J
12. Hegarty WH, Sims H. Some determinants of unethical Pers. 1989;3:107-19.
decision behavior: An experiment. J Appl Psychol. 31. Walters GD. The Criminal Lifestyle: Patterns of Serious
1978;58:111-21. Criminal Conduct. Newberry Park, CA:Sage; 1990.
13. Victor BC, Trevino K, Shapiro DL. The influence of 32. Emmons RA. Relationship between narcissism and
justice evaluations and social context factors. J sensation seeking. Psychol Rep. 1981;48:247-50.
Business Ethics. 1993;12:253-63. 33. Riskind JH, Kelly K, Harmen W, Moore R, Gaines HS.
14. Darley JM. Social organization for the production of The loomingness of danger: Does it discriminate all
evil. Psychol Inquiry. 1992;3:199-218. phobia and general anxiety from depressing? Cognitive
15. Ones DS, Viswesvaran C, Schmidt FL. Comprehensive Ther Res. 1992;16:602-22.
meta-analysis of integrity test validities: Findings and 34. Mumford MD, Connelly MS, Helton WB, Mowry JR,
implications for personnel selection and theories of job Osburn HK. On the construct validity of integrity tests:
performance. J Appl Psychol Monograph. Individual and situational factors as predictors of test
1993;78:674-703. performance. Int J Selection Assessment. [In press].
16. Sackett PR, Waneck JE. New developments in the use 35. Ash P. Predicting dishonesty with the Reid Report. J
of measures of honesty, integrity, conscientious, Am Polygraph Assoc. 1985;5:139-45.
dependability, trustworthiness, and reliability for 36. Harris MM, Sackett PR. A factor analysis of an
personnel selection. Personnel Psychol. 1996;49:787- employee honesty test. J Business Psychol.
829. 1987;2:122-35.
17. Hogan J, Brinkmeyer K. Bridging the gap between 37. Gough HG. The California Psychological Inventory.
overt and personality-based integrity tests. Personnel In: Newmark CS, ed. Major Psychological Assessment
Psychol. 1997;50:587-600. Instruments: Volume II. Boston, MA: Allyn and Bacon;
18. Trevino L, Youngblood SA. Bad apples in bad barrels: 1989:67-98.
A causal analysis of ethical decision making behavior. J
Appl Psychol. 1990;75:378-85.
87
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
38. Mumford MD, Costanza DP, Connelly MS, Johnson JF. 55. Amabile TM. Motivational synergy: Toward new
Item generation procedures and background data scales: conceptualizations of intrinsic and extrinsic motivation
Implications for construct and criterion-related validity. in the work place. Hum Resource Management Rev.
Personnel Psychol. 1996;49:361-98. 1993:3;185-201.
39. Gessner TE, O’Connor JA, Mumford MD, Clifton TC, 56. Collins MA, Amabile TM. Motivation and creativity.
Smith JA. Situational variables influencing the In: Sternberg RJ, ed. Handbook of Creativity.
propensity for destructive acts: Taxonomy development Cambridge, England: Cambridge University Press;
and validation. Curr Psychol. 1995;13:303-25. 1999:297-312.
40. Wild JT. The origin of soft-tissue ultrasonic echoing 57. Bowie NE. A Kantian theory of meaningful work. J
and early instrumental application to clinical medicine. Business Ethics. 1998;17:1083-92.
In: Weber RJ, Perkins DN, eds. Inventive Minds: 58. Eisenberger R, Cammeron J. Detrimental effects of
Creativity in Technology. Cambridge, England: reward: Reality or myth? Am Psychol. 1996;51:1153-6.
Cambridge University Press; 1992. 59. Baumhart R. How ethical are businessman. Harvard
41. Gruber HE, Wallace DB. The case study method and Business Rev. 1961;39:6-19.
evolving systems approach for understanding unique 60. Bellizi JA, Hite RE. Supervising unethical sales force
creative people at work. In: Sternberg RJ, ed. behavior. J Marketing. 1989;53:36-47.
Handbook of Creativity. Cambridge, England: 61. Deconinck JB. How sales managers control unethical
Cambridge University Press; 1999:93-115. sales force behavior. J Business Ethics. 1992;11:789-
42. Vosberg SK. The effects of positive and negative affect 98.
on divergent thinking performances. Creativity Res J. 62. Harrison JS, Fiet JO. New CEOs pursue their own self-
1998;11:165-72. interests by sacrificing stakeholder value. J Business
43. Fiedler FE, Garcia JE. New Approaches to Effective Ethics. 1999;19:301-8.
Leadership: Cognitive Resources and Organizational 63. Pelz DC, Andrews FM. Scientists in Organizations:
Performance. New York: Wiley; 1987. Productive Climates for Research and Development.
44. Chang MK. Predicting unethical behavior: A New York: Wiley; 1966.
comparison of the theory of reasoned action and the 64. Simonton DK. Historiometric studies of creative
theory of planned behavior. J Business Ethics. genius. In: Runco MA, ed. The Creativity Research
1998;17:1825-34. Handbook: Volume One. Cresskill, NJ: Hampton Press;
45. Goolsby JR, Hunt SD. Cognitive moral development 1997:3-28.
and marketing. J Marketing. 1992;56:55-68. 65. Mullins RF, Sherman R. Creativity and performance
46. Weeks WA, Moore CW, McKinney JA, Longenecker appraisal: Shall never the twain meet. Creativity Res J.
JG. The effects of gender and career stage on ethical 1993;6:425-34.
judgement. J Business Ethics. 1999;20:301-13. 66. Yukl G. Leadership in Organizations. Englewood
47. Arlow P, Ulrich TA. Business ethics and social Cliffs, NJ: Prentice-Hall; 1998.
responsibility of business students: An empirical 67. Schminke M, Wells D. Group processes and
comparison of Clark’s study. Akron Business Economic performance and their effects on individual’s ethical
Rev. 1980;11:17-23. frameworks. J Business Ethics. 1999;18:367-81.
48. Chonko LB, Hunt SB. Ethics and marketing 68. Zabid ARM, Alasgoff, SK. Perceived ethical values of
management: An empirical examination. J Business Malaysian managers. J Business Ethics. 1993;12:331-
Res. 1985;13:339-59. 7.
49. Kidwell JM, Stevens RC, Bethke RL. Differences in the 69. Sims RR. Changing an organization’s culture under
ethical perceptions between male and female managers: new leadership. J Business Ethics. 2000;25:65-78.
Myth or reality. J Business Ethics. 1987;6:489-93. 70. Minkes AC, Small MW, Chatterjee SR. Leadership and
50. Teal EJ, Carroll AB. Moral reasoning skills: Are business: Does it matter? Implications for management.
entrepeneurs different. J Business Ethics. 1999;19:229- J Business Ethics. 1999;20:327-35.
40. 71. House RJ, Howell JM. Personality and charismatic
51. Reiss MC, Mitra K. The effects of individual difference leadership. Leadership Q. 1992;3:81-108.
factors on the acceptability of ethical and unethical 72. Meyer AG, Brown T, Browne MN, Kubasek N. Do we
workplace behaviors. J Business Ethics. 1998;17:1581- really want more leaders in business? J Business Ethics.
93. 1998;17:1727.
52. Feist GJ, Gorman ME. The psychology of science: 73. Vredenburgh D, Brender Y. The hierarchical abuse of
Review and integration of a nascent discipline. Rev power in work organizations. J Business Ethics.
Gen Psychol. 1998;2:3-47. 1998;17:1337-1347.
53. Feldman DH. The development of creativity. In: 74. Baron F, Harrington D. Creativity, intelligence, and
Sternberg RJ, ed. Handbook of Creativity. Cambridge, personality. Annu Rev Psychol. 1981;32:439-76.
England: Cambridge University Press; 1999:169-88. 75. Mumford MD, Whetzel DC, Reiter-Palmon R.
54. Csikszentmihalyi M. Implications of a system’s Thinking creatively at work: Organizational influences
perspective for the study of creativity. In: Sternberg RJ, on creative problem solving. J Creative Behav.
ed. Handbook of Creativity. Cambridge, England: 1997;31:7-17.
Cambridge University Press; 1996:313-38.

88
––––––––––––––––––––––––––––––––––––––––––––––– Mumford & Helton, Organizational Influences on Scientific Integrity
76. Houndshell DA. Invention in the industrial research 93. Castrogiovani GT. Pre-start-up planning and the
laboratory: Individual act on collective process? In: survival of new small businesses: Theoretical linkages.
Weber RN, Perkins DN, eds. Inventive Minds: J Management. 1996;22:801-22.
Creativity in Technology. Cambridge, England: 94. Morris MH, Marks AS, Allen JA, Perry NS. Modeling
Cambridge University Press; 1992:273-390. ethical latitudes and behaviors under conditions of
77. Reall M, Bailey JJ, Stoll SK. Moral reasoning on hold environmental turbulence: The case of South Africa. J
during a competitive game. J Business Ethics. Business Ethics. 1996;15:1119-30.
1998;17:1205-10. 95. Roussouw GT. Establishing moral business culture in
78. Schwepker CH. Understanding salespeople’s intention newly formed democracies. J Business Ethics.
to behave unethically: The effects of perceived 1998;17:1563-71.
competitive intensity, cognitive moral developments, 96. Verschoor CC. A study of the link between a
and model judgement. J Business Ethics. 1999;21:303- corporation’s financial performance and its commitment
16. to ethics. J Business Ethics. 1998;17:1509-16.
79. Feist GJ. The influence of personality on artistic and 97. Judge WO. Correlates of organizational effectiveness: A
scientific creativity. In: Sternberg RJ, ed. Handbook of multilevel analysis of a multi-dimensional outcome. J
Creativity. Cambridge, England: Cambridge University Business Ethics. 1994;13:1-10.
Press; 1999:273-96. 98. Zarkada-Fraser A. A classification of factors
80. Zuckerman H, Cole JR. Research strategies in science: influencing participating in collusive tendering
A preliminary inquiry. Creativity Res J. 1994;7:391- agreements. J Business Ethics. 2000;23:269-82.
406. 99. Enderle G, Tavis LA. A balanced concept of the firm
81. Pelton LE, Chowdhury J, Vitell SJ. A framework for the and the measurement of its long-term
examination of relational ethics: An interactionist planning and performance. J Business Ethics.
perspective. J Business Ethics. 1999;19:241-53. 1998;17:1124-44.
82. Rallapalli KC, Vitell SJ,, Szeinbach S. Marketers’ 100.Koehn D. Can and should businesses be friends with
norms and personal values: An empirical study of one another and their stakeholders. J Business Ethics.
marketing professionals. J Business Ethics. 1998;17:1755-63.
2000;24:65-75. 101.Valenzuela JLD, Villaconta FS. The relationship
83. Cote J, Goodstein J. A breed apart? Aecurity analysis between companies and their suppliers. J Business
and herding behavior. J Business Ethics. 1999;18:305- Ethics. 1999;22:273-80.
14. 102.Jones MT. The institutional determinants of social
84. Gessner TE, O’Connor JA, Clifton TC, Connelly MS, responsibility. J Business Ethics. 1999;20:163-79.
Mumford MD. The development of moral beliefs: A 103.Vitell S, Festervand TA. Business ethics: Conflicts,
retrospective study. Curr Psychol. 1993;11:236-59. practices, and beliefs of industrial executives. J
85. Leede JD, Nijhof AHJ, Fisscher OAM. The myth of Business Ethics. 1987;6:111-22.
self-managing teams: A reflection on the allocation of 104.Dubinsky AJ, Ingram TN. Correlates of salespeople’s
responsibilities between individuals, teams, and the ethical conflict: An exploratory investigation. J
organization. J Business Ethics. 1999;21:203-15. Business Ethics. 1984;3:343-53.
86. Barnett T. A preliminary investigation of the 105.Dooley RS, Fryxell GE. Are conglomerates less
relationship between selected organizational environmentally responsible? An empirical examination
characteristics and external whistle blowing by of diversification strategy and subsidiary pollution in the
employees. J Business Ethics. 1992;11:949-59. U.S. chemical industry. J Business Ethics. 1999;21:1-
87. Kawathatzopoulos I. Development of cognitive skill in 14.
solving business ethics problems: The effect of 106.Reichers AE, Schneider B. Climate and culture: An
instruction. J Business Ethics. 1993;12:379-86. evolution of constructs. In: Schneider B, ed.
88. Verbke WC, Ouwerkerk C, Peelen E. Exploring Organizational Climate and Culture. San Francisco,
contextual and individual factors on ethical decision- CA: Josey-Bass; 1990:5-39.
making of sales people. J Business Ethics. 107.Sims RL, Keon TL. Determinants of ethical decision
1996;15:1175-87. making: The relationship of the perceived organizational
89. Weaver KM, Farrell OC. The impact of corporate environment. J Business Ethics. 1999;20:393-404.
policy in reported ethical beliefs and behavior of 108.Agarwal J, Malloy DC. Ethical work climate decisions
marketing practicioners. AMA Proc. 1977;27:477-81. in a not-for-profit organization: An empirical study. J
90. Fritz JMH, Arnett RC, Conkel M. Organizational Business Ethics. 1999;20:1-14.
ethical standards and ethical commitment. J Business 109.Vidader-Cohen D. Moral climate in business firms: A
Ethics. 1999;20:289-92. conceptual framework for the analysis of change. J
91. Grimalda G. Participation versus social exclusion. J Business Ethics. 1998;17:1211-26.
Business Ethics. 1999;21:269-79. 110.Key S. Organizational ethical culture: Real or
92. Schokkaert E, Sweeney J. Social exclusion and ethical imagined? J Business Ethics. 1999;20:217-25.
responsibility: Solidarity with the least skilled. J 111.Argandona A. Sharing out in alliances: Trust and ethics.
Business Ethics. 1994;21:251-67. J Business Ethics. 1999;21:217-28.

89
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
112.Hartman SJ, Yrle AC, Galle WP. Procedural and
distributive justice: Examining equity in a university
setting. J Business Ethics. 1999;20:337-51.
113.Amabile TM, Griskiewicz N. The creative environment
scale: The work environment inventory. Creativity Res
J. 1989;2:231-54.
114.Gotterbarn D. Not all codes are created equal: The
software engineering code of ethics, a success story. J
Business Ethics. 1999;22:81-9.
115.Etheredge JM. The perceived role of ethics and social
responsibility: An alternative scale structure. J Business
Ethics. 1999;18:51-64.
116.Brierley JA, Cowton CJ. Putting meta-analysis to work:
Accountants’ organizational-professional conflict. J
Business Ethics. 2000;24:343-53.

90
3. Medical Practice and Clinical Research
Waiving Informed Consent: Long-Term Consequences for the U.S. Military
Mary L. Cummings, Engineering Fundamentals, Virginia Tech, USA

Keywords: Anthrax, Gulf War, Informed consent, Investigational drugs, Military, Waiver

In December 1990, the Department of Defense (DoD), anticipating the invasion of Kuwait for
Operation Desert Storm, petitioned the Federal Drug Administration (FDA) to waive the federally
mandated informed-consent requirements in the case of two investigational drugs: pyridostigmine
bromide (PB) and botulinum toxoid (BT). PB, administered orally, was thought to be an effective
pre-treatment against the nerve agent soman. The BT vaccine was potentially effective against the
bacterium causing botulism (1). Fearful of the possibility that Saddam Hussein would conduct
chemical and biological warfare against American troops, the Joint Chiefs of Staff felt that these two
investigational drugs could protect U.S. soldiers. The concerns of military leadership were well-
founded. Saddam Hussein had used chemical nerve agents and mustard gas against his own people in
the Iran-Iraq War (2). However, while military intelligence confirmed that Iraq had the capability to
make biological and chemical (nerve agent) weapons, no evidence indicated Iraq had ever made a
weapon with soman (3).
FDA did not approve PB and BT. They were considered experimental and fell under the category
of investigational new drug (IND). Federal regulations stipulate that if any Federal agency, including
the military, desires to use an unapproved drug, that agency must first fully brief the individuals
receiving the IND. This briefing must include mention of associated drug use hazards, and the
potential recipients’ written consent must be obtained. Prior to the Gulf War, informed consent for
INDs could only be waived in extreme emergencies, even for the military. However, the U.S. military
determined that it was not feasible to seek the informed consent of 700,000 personnel deployed to the
Middle East. In 1990, in the months preceding the Gulf War, the military petitioned the FDA to
waive the informed consent regulations. The FDA, not wishing to intervene in national security
policy and with the approval of an Institutional Review Board (IRB), issued the waiver in an interim
ruling in December 1990 (4). However, as part of the approval for the waiver, the military was
required to provide information sheets about PB and BT to the recipients detailing the possible side
effects. In addition, the military was expected to carefully document the use of the INDs as well as
any adverse reactions.
Approximately 300,000 military personnel received the PB pills and 8000 individuals received
the BT vaccine during the Gulf War (5). Despite the specific requirement by the FDA that the
military track data on both drugs, no procedure was ever established to document which personnel
received the drugs and if any adverse side effects were noted (1). Many military personnel
experienced systemic medical problems both during and after the Gulf War that were not combat
related. Such problems have been termed as the Gulf War Syndrome (GWS). Most notably, over
100,000 Gulf War veterans complained of maladies ranging from chronic fatigue to paralysis in the

Corresponding author: Mary L. Cummings, 332 Randolph Hall,Virginia Polytechnic Institute and State University,
Blacksburg, VA 24061, 540-231-6555 (voice), 540-231-6903 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
years immediately following the war (3), and of (9).
these, 20,000 reported debilitating symptoms (6). Even though FDA waived the requirement
In preliminary studies, PB has now been for obtaining informed consent for the use of PB
implicated as the primary catalyst of the GWS, and BT in the Gulf War, the approval was
however the research is still in its early stages contingent upon the military providing those
(3). service members who received the INDs with
information sheets describing the PB and BT
Waiving Informed Consent treatments in detail. The sheets were to explain
The Federal regulations that govern informed the reasons for using the INDs, the symptoms of
consent for human subjects fall under the botulism and a nerve agent attack, and most
purview of the Department of Health and Human importantly any potential side effects or
Services (DHHS). The regulations state that reactions. In addition, the soldiers were also
informed consent may be waived when using asked to report any of these side effects or
INDs, but a number of conditions must be met. reactions. Apparently, the information sheets
No more than minimal risk can exist for the never made it to the Gulf War theater, so the
patient, and after the treatment is concluded, the personnel who received the treatments did not
participants must be notified of both the receive any written information about the INDs.
procedure and the possible risks (7). FDA, bound However, even a cursory glance at the
by the DHHS regulations, established their own information sheets that were approved by the
framework of rules regarding INDs. Prior to the Army for dissemination shows that they were at
Gulf War waiver, FDA maintained that the best superficial.
informed consent process could be waived only
in a life-threatening emergency with the patient Ethical Issues
unable to communicate and without time to In 1978, the National Commission for the
obtain consent from patient’s legal representative Protection of Human Subjects of Biomedical and
(7). Behavioral Research issued the Belmont Report
The Joint Chiefs of Staff decided it was not that identified three principles that are
feasible to obtain the informed consent of fundamental in determining whether a research
700,000 military personnel deployed to the Gulf protocol is ethical. They are: respect for persons,
War region and that the pending conflict was beneficence, and justice. These are the primary
essentially an emergency situation by FDA ethical considerations of an IRB when evaluating
standards. However, prior to granting the a research protocol (10). The crux of the respect-
military informed consent waivers for the use of for-persons principle is the preservation of a
PB and BT, FDA required the military to convene person’s autonomy when making decisions about
an IRB (1). To meet this Federal requirement for his/her own medical care. It is this aspect of the
the BT vaccine, the military actually convened Belmont Report that is at issue in waiving
two IRBs. The first IRB, the U.S. Army Medical informed consent. By swearing an oath to the
Research Institute of Infectious Diseases military and the nation, service members
(USAMRIID) Human Use Committee, was the willingly sacrifice some autonomy concerning
panel typically used by Army research personnel decisions about their own lives. Enlisting in the
to consider protocols involving human subjects. military is a supreme sacrifice and highly
The USAMRIID concluded that it was unethical commendable, but should soldiers lose all rights
to waive the informed consent of military to autonomy, especially when it comes to their
personnel who would receive BT (8). They health? The DoD defends its actions in waiving
further recommended that oral, not written, informed consent for INDs by stating, “Allowing
consent be obtained because oral consent was a soldier to refuse treatment would endanger him/
feasible, and it also respected the rights of the her as well as those who would try to save their
soldiers. Six days later, for reasons not stated in lives and ruin mission success”(5). This
any DoD documents or in any IRB minutes, the paternalistic approach by the DoD overlooks one
DoD then convened a second, entirely different critical aspect: What exactly constitutes
IRB, the Surgeon General’s Human Subjects “treatment?”
Research Review Board (HSRRB). The HSRRB There has been much debate as to whether
approved the BT protocol as submitted and the military’s use of PB and BT constitutes
recommended that informed consent be waived research or treatment. In the clinical trials held
94
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Cummings, Waiving Informed Consent
months before the Gulf War, only a select group with no impediments. The military was clearly
of male human subjects were tested with PB and circumventing the system and in doing so
BT. There was no testing for interactions with trivialized the IRB process and violated Federal
other chemicals or drugs likely to be used with regulations. It appears the military was only
the INDs, and no long-term studies were seeking IRB approval as a formality in an
conducted (5). Additionally, persons with health administrative procedure and lost sight of the
problems typical of military populations were purpose of the review. FDA, very concerned
never studied in conjunction with the drug about the military’s use of multiple IRBs when
testing, and women never participated in any seeking informed consent waivers, censured the
trials (2). Is it ethical and reasonable to maintain military in October of 1999 for this violation and
that military members receiving drugs tested on a changed the federal regulations regarding
very small, isolated population were receiving military IRBs (1). As a result, IRBs convened by
“treatment?” Despite the fine line between the military to evaluate IND protocols are now
treatment and research with investigational required to include at least three members who
drugs, FDA’s own regulations clearly state that are not employees or officers of the federal
informed consent is required even when the government and are not affiliated with the
unapproved drug is to be used in a therapeutic protocol in any way.
manner because the drug has not yet passed full
FDA efficacy and safety trials (11). Long-Term Consequences
The respect-for-persons principle was again In December 1997, DoD announced plans to
violated when the information sheets for the vaccinate all 2.4 million U.S. troops against the
INDs were “lost” (5, 12). These sheets should biological threat of anthrax. If not treated in its
have been paramount in the minds of military initial stages, anthrax is deadly (13). The current
medical professionals overseeing the PB & BT anthrax vaccine is approved by the FDA and was
programs. The IRB approval and FDA originally designed for agricultural workers and
authorization for PB and BT were contingent on veterinarians. It is a six-shot protocol that is
the investigators adhering to the approved administered over a period of 18 months.
protocols, which included the distribution of the Because of this extended treatment period, DoD
information sheets. The INDs found their way decided that it must vaccinate all 2.4 million
successfully to the Gulf War theater, and if DoD personnel in the unlikely event that all U.S.
leadership had considered the sheets a similar forces faced a biological threat.
priority, they would have been delivered also. Almost immediately after DoD made its
Did the military view the information sheets as announcement, military members began to
“not feasible” just as they did for informed protest, based in part on the revelation that
consent? When FDA later evaluated the service members were given experimental drugs
military’s use of INDs during the Gulf War, it without their knowledge in the Gulf War.
identified “significant deviations from Federal Military, medical, and legal critics of the anthrax-
regulations published in Title 21, Code of Federal vaccine decision were not satisfied that the
Regulations (CFR), parts 50 and 312.” (1). FDA vaccine was approved by the FDA (13 -15). The
cited several areas in which the military was not sole manufacturer of the anthrax vaccine,
in compliance. Most notably FDA admonished Michigan Biologic Products Institute (now Bio-
the military for not disseminating the information Port) has failed numerous FDA inspections.
sheets prior to the use of INDs in the Gulf War. Most recently, Bio-Port was cited for 23
FDA also issued DoD a stern reprimand for not violations, some of which included sterility and
keeping detailed records on who received the potency deviations, and some microbial
drugs and, most importantly, any adverse contamination (14, 15). In fact, to date the
reactions suffered by military personnel. Michigan plant still has not passed an FDA
Lastly, the most glaring ethical issue was inspection (15, 16).
DoD’s use of two different IRBs. When the There have never been any published studies
Army’s first IRB found that it was unethical to of human efficacy or long-term effects for the
administer BT to military personnel without their anthrax vaccine (15). Moreover, according to an
informed consent, the DoD convened a second April 1999 General Accounting Office (GAO)
IRB that produced the desired result of report, long-term effects of the anthrax vaccine
recommending the waiver of informed consent have never been studied. To further add to the
95
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
debate over the efficacy of the anthrax vaccine, data, the biological warfare threat for U.S. troops
the Institute on Medicine has stated that the has not changed since 1990 (14).
licensed anthrax vaccine is only effective against
cutaneous anthrax and furthermore has never A Final Note on Accountability
been tested for pulmonary anthrax, which would Accountability is an imperative moral trait
be the method of delivery in a combat arena (13). required of all military personnel and is
A chief Army biological researcher wrote in a considered the cornerstone for military command
1994 textbook on vaccines that “the current and leadership. By court-martialing military
vaccine against anthrax is unsatisfactory” (14). personnel who refuse the anthrax vaccine, DoD
Despite the military’s assertions that it is only is holding these people accountable for their
interested in protecting the welfare of its soldiers, actions. For those court-martialed, this
GAO charges that DoD is extremely negligent in accountability will not cost them just their jobs
tracking adverse reactions to the anthrax vaccine, within the military. In addition, they are
which was a significant problem with the INDs dishonorably discharged and lose all their
used in the Gulf War. In fact, many military veterans’ benefits as well as their retirement
personnel have reported adverse reactions to the benefits. The nation recognizes the right to make
anthrax vaccine. However, in the absence of any autonomous health-related decisions for all
established tracking and monitoring system, there citizens, but it appears, not for military personnel
is no way to accurately identify any percentages. who pay a high price for both autonomy and
With the data supporting the questionable accountability.
status of the anthrax vaccine and considering This exacting level of military discipline and
DoD’s past history, it is not unreasonable to accountability is unfortunately glaringly absent
expect military personnel to have doubts about from DoD’s use of INDs in the Gulf War.
both the efficacy of the anthrax vaccine and the Especially troubling are the following:
military’s plans for implementation. To combat • DoD convened a second IRB for an IND
potential insubordination, DoD court-martialed protocol when the first did not produce the
those personnel who refused the vaccine, stating desired recommendation to waive informed
that allowing soldiers to refuse the vaccine would
undermine discipline and be prejudicial to good consent.
order. Many military members, outraged at • No one was held accountable for the lost
DoD’s response and facing involuntary information sheets in the Gulf War. If
inoculation, chose to resign from the service military officers lost strategic documents
rather than risk their health. The military is protecting troops’ safety, they would most
already facing serious retention and recruiting definitely face severe punishment.
problems, and DoD’s refusal to make the anthrax • No one was held accountable for the incred-
vaccine voluntary is only adding to an already ible lack of record keeping including track-
critical personnel shortage. ing adverse reactions during and after the
Prior to the mandated anthrax vaccination of Gulf War. Not only did military personnel
all U.S. troops, the military’s policies against the
threat of chemical and biological warfare were suffer from a lack of treatment information,
deterrence, containment of the enemy, and use of but also the entire medical field suffered
other defensive measures such as protective suits from the loss of critical data.
and warning devices (13). It was not until the This clear double standard in accountability will
Gulf War that troops were inoculated against the only continue to haunt the military. Public
threat of possible biological warfare, and it was reports on the military’s use of experimental
not until 1997 that troops were forcibly drugs on troops without their knowledge and the
inoculated in peacetime. There has been much anthrax debacle will only continue to exacerbate
criticism directed toward DoD for implementing personnel issues. FDA has recently issued more
the anthrax vaccine in peacetime. DoD stringent rulings to prevent some of these ethical
responded that even though there is no threat of transgressions from occurring in the future and to
war, the 18-month treatment period for the compel the military to abide by the laws they are
anthrax vaccine requires that it must prepare its supposedly defending. However, not until DoD
forces for any future contingencies. However, embraces the Federal policies designed to respect
GAO asserts that based on military intelligence basic human rights and autonomy will the
96
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Cummings, Waiving Informed Consent
military regain some of its medical credibility
and confidence in leadership.

Acknowledgments
I would like to extend a special thanks to Dr.
Doris T. Zallen for her guidance and technical
support.

Bibliography
1. 64 Federal Register at 54,180; October 5, 1999.
2. Milner CA. Gulf War guinea pigs: Is informed consent
optional during war? The Catholic University of
America Journal of Contemporary Health Law & Policy
1996 Fall; 13: 199-232.
3. Myers SL. Drug May Be Cause of Veterans’ Illnesses.
New York Times 1999 Oct 18; Sect. A:18, (col. 4).
4. 55 Federal Register at 52,187 (1990) codified at 21 CFR
50.23(d).
5. Ryan RP. Should combat troops be given the option of
refusing investigational drug treatment? The Food and
Drug Law Institute Food and Drug Law Journal 1997;
52(4): 377-400.
6. Tuthill KA. Human Experimentation: Protecting Patient
Autonomy Through Informed Consent. The Journal of
Legal Medicine 1997 Jun; 18: 221-250.
7. Code of Federal Regulations: 45 CFR 46.116(c), (1999).
8. U.S. Army Medical Research Institute of Infectious
Diseases memorandum SGRD-UIZ-H. Fort Detrick,
Maryland: Department of the Army; October 5, 1990.
9. Office of the Surgeon General memorandum SGRD-HR
(15-1a). Falls Church, VA: Department of the Army;
October 11, 1990.
10. The Belmont Report: Ethical Principles and Guidelines
for the Protection of Human Subjects of Research.
Washington DC: DHEW Publication (OS) 78-0011;
1978.
11. Code of Federal Regulations: 21 CFR 312.35 (2000).
12. The Pill. “60 Minutes.” CBS television; September 29,
1996.
13. Havrilak GT. The Pentagon’s Anthrax Vaccination
Immunization Program. Military Readiness Project;
1999 Dec.
14. Rempfer TL. Why Am I Resisting the Vaccine? The
Military Trained Me To. The Washington Post 2000 Jan
3; Sect. B:01.
15. Meryl Nass MD. Written testimony before the
Subcommittee on National Security, Veterans’ Affairs,
and International Relations, U.S. House of
Representatives. April 19, 1999.
16. The Court-Martial of Major Sonnie Bates: “60
Minutes.” CBS television; February 6, 2000.

97
Promoting Scientific Integrity: The Long Road Ahead–Some Considerations
from Espírito Santo, Brazil
Jaime Roy Doxsey, Department of Social Sciences, Federal University of Espírito Santo,
Brazil

Keywords: Brazil, Research integrity, Social and human sciences, Teaching research ethics

“We live in an historical moment of transformation of the scientific paradigm which

questions the criteria that scientific rigor in and of itself is ethical.” BB Sawaia, 1999.

While the promotion of research integrity has tended to receive widespread governmental and
institutional support in the United States and Canada, the responsible conduct of research, including
preventing and handling of misconduct, are not always prominent issues in many developing
countries such as Brazil. This paper examines the need to stimulate institutional awareness and
debate on major issues such as production and communication of scientific knowledge as well as the
ethical challenges for developing responsible research practices in the human and social sciences.
A lack of Federal or state legislation, institutional policies or public concern regarding the quality
and the ethics of scientific research do not exempt researchers or universities from establishing
programs to insure research integrity. The institutional context of a medium-sized Federal
government university, the Federal University of Espírito Santo, is examined in an attempt to describe
work conditions, the institutional culture and other obstacles for establishing a program to promote
research integrity.
In Brazil, recent Federal resolutions in the areas of health, medicine and medical research have
established guidelines for human protocol, research integrity, and the protection of human subjects
and have determined a local project review procedure along the lines of North American legislation.
These guidelines extend themselves to all scientific or academic research activities that involve
human subjects. The Brazilian university system and the National Council for Research (CNPQ),
however, have neither acknowledged the relevance of these resolutions for research practices nor
incorporated them into grant procedures.
At the local level, universities, research institutes, academic centers, departments and graduate
programs establish their own policies for research projects and scientific production. Institutional
procedures seldom exist for handling allegations of scientific misconduct or establishing protocols for
human subjects.
The recent expansion of the number of graduate programs also has increased the need for
programs to promote the teaching of research integrity, the ethics of mentoring, and academic career
pressures. Further, data management, recording, retention, etc., require pro-active policies to
anticipate conflicts and incidents of misconduct.
Corresponding author: Jaime R. Doxsey, Department of Social Sciences, Federal University of Espírito Santo, Rua
Pernambuco, 81/1102, Praia da Costa, Vila Velha, Espírito Santo 29.101-335, Brasil, 55-27-3349-8992 (voice), 55-27-3335-
7614 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
What are the implications of these conditions Rectory for Graduate Study and Research,
for research with human subjects in Brazil? Is Federal University of Espírito Santo and the
the Brazilian population unduly exposed to annual reports of the Office of Research Integrity,
doubtful research practices and scientific Department of Health and Human Services, U.S.
misconduct, particularly the lower population Office of Public Health and Science. The journal
strata (over 50% of the total population) and editions of the Cadernos de Ética em Pesquisa
more specifically, vulnerable sectors of this [Notebooks of Research Ethics], published by the
population? Brazilian National Commission of Research
At first glance, the answer would be an Ethics were specially useful in providing
uncategorical “no”. Even considering the lack of background information for this text.
a more systematic analysis of actual research
practices, there is no direct or indirect evidence Results–The Brazilian Context
that medical, health, human, or social sciences In Brazil, Federal resolutions first established the
research in Brazil is unethical. What could be National Commission of Research Ethics
considered unethical is the lack of priority for (CONEP) in 1996 and determined guidelines for
such research at all levels of government in light human protocol, research integrity, and the
of the rising indices of preventable social protection of human subjects in 1997. The 1997
diseases, human violence, drug abuse, and the resolution determined a project review procedure
subsequent decline of living conditions/quality of in the areas of health, medicine, and medical
public services for the lower strata of the research by local Committees of Ethics and
population. Research. At the present time, there are
With financial support and investment in approximately 266 Committees of Ethics and
social policies at an astonishingly low level, Research (CEPs), the majority of which are
social research tends to be descriptive, located in institutions related to medical
exploratory, or action-oriented. Academic instruction or university-associated hospitals.
research seldom receives external or internal Although the guidelines extended themselves
financing, and most funding is limited to to all scientific or academic research activities
scholarships for undergraduate trainees or the that involve human subjects, the Federal
support of field work. Brazilian university system and the CNPQ have
The lack of a regulatory system of project neither acknowledged the relevance of these
approval and norms for the protection of human resolutions for research practices nor
subjects should not be misinterpreted as a lack of incorporated them into institutional procedures.
research ethics. In a country like Brazil, the few Data from CONEP reveal the registration of
individuals actively engaged in research with 559 projects in 1999. In a classification by
human subjects do so with great dedication and Specialty Topics, most of these projects were
considerable respect for their human subjects. grouped under the topic of “international
Ethical values are not necessarily culturally cooperation” (78.3%), and a majority within this
ascribed or limited by adverse institutional and category (80%) involved new medications.
social conditions. Distribution in other topical areas included
Nevertheless, what are the actual human genetics (7.8%), reproduction (5%),
circumstances in which the social and human indigenous populations (1.6%), new medical
sciences are being practiced in Brazil? In what procedures, and equipment (5.3%) (1).
institutional context might it be necessary to In observance of the data cited above, it is
initiate the promotion of research integrity and at not surprising to conclude that medical and
least provide guidelines for misconduct health research formally lead the way in
regulation? How may this promotion of research establishing human protocols for research with
integrity be best approached? human subjects. Also, it is not accidental that the
majority of the projects reviewed involve
Design international funding and/or cooperative
This paper is a descriptive essay based on agreements. A recent review of the literature
personal observations and a review of scientific available within Brazil points exclusively toward
journals, research methodology textbooks bioethics and medical and health ethics as
published in Portuguese, Internet homepages, dominant topics in the field of ethical
records of research projects available in the Pró- considerations (2).
100
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Doxsey, Promoting Scientific Integrity
In the humans sciences, there is little to focused on the role of the journals in improving
report. However, in 1997, the Federal Council of research quality, technical aspects of the journals,
Psychology determined that new methods or and proceedings for evaluation/selection of
procedures in the field could be utilized if articles. The last session included an item on
presented as research following research norms scientific and ethical aspects of journal editing.
for human subjects. The Committee of Ethics in Increased public concern with electoral
Research at the Catholic University of São Paulo opinion polling has attracted attention in the last
(Catholic University – SP) was implemented national elections for president and congress, and
through the work of a sociologist who lead most recently in municipal elections. The
discussions to delimitate general principles concern voiced by media and politicians is
regarding research ethics, which “took into directed, however, to the possible undue
consideration the specificity, plurality and influence of the poll results on the voter and the
scientific creativity of the production of political system. No ethical concern for poll
knowledge in the human sciences” (3). subjects has been registered. Issues regarding
Unlike the CEPs created in the medical area, informed consent, the use of the poll results, or
at the Catholic University-SP, the Committee has the subjects’ knowledge of the funding sources
developed educational functions to represent the have not been publicly evaluated.
ethical principles of the institution, serving as a Although the lack of governmental support
review board for special recourses. Research for scientific and technological research and
projects that are considered to have special development is a constant criticism throughout
ethical questions are sent to the Committee by the Brazilian society, there is no strong public
academic orientators, or by dissertation, thesis, or support for financing academic research.
research commissions for educational Resources from private and international
evaluations. This university understood that foundations are centered on corporate interests
ethical evaluations were already occurring at with little direct university participation. In
other institutional levels and that the short, there is little grant money, private or
centralization of the approval process in one public, which might warrant an institutional
committee would be not only impossible but policy being created in order to qualify for grant
would fail to capture the different optics of applications.
research ethics. While international funding or “cooperation”
Another indicator of the extent of concern for might be instrumental in aligning research
research integrity was presented in a study interests in the biomedical sciences to installing
entitled: “Analysis of ethical aspects of research parallel regulatory proceedings for research
in human beings contained in the authors’ ethics, there are no similar external stimuli for
instructions of 139 Brazilian scientific journals”. the human and social sciences in Brazil. With no
(4) Although the study was limited to a review public pressure or support for human research,
of scientific journals in the areas of medicine, little or no funding, and a lack of issues that
nursing, odontology, and the general sciences, the might stimulate institutional response tend to
authors discovered that 79% of the journals made neutralize the need for more relevant,
no reference to ethical considerations in their modernized research policies in the Brazilian
notes to potential contributors. Only 12% of the University system.
journals made reference to the necessity of
approval or analysis of the research project by a A Short Case Study–the UFES
Committee or Commission of Ethics in Research. Current research policies at the Federal
This author has no knowledge of instructions University of Espírito Santo deal principally with
to authors in the area of the social and human the administrative approval of faculty
sciences. With the growing number of scientific involvement in research as well as release time
publications in Brazilian universities, there is from academic classroom schedules.
some concern for establishing selection processes Authorization to conduct research is granted by
for articles and the evaluation process of the the department council, after a written evaluation
journals. During May, the Faculty of Education often by a research commission of peers. A
at the University of São Paulo organized a simplified regulatory system presently requires
conference to discuss the publication policies of project approval by the council of department
scientific journals in education. Discussion was heads at the level of the academic center and
101
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
eventual registration of the project in the Pró- involving misconduct in research, plagiarism,
Rectory for Graduate Studies and Research. misrepresentation of academic production or
Details of the project must be outlined on a other problems of research integrity can only be
basic form that specifies the usual information handled administratively under the existing
regarding the nature of the study, authors, legislation and institutional procedures (5).
methods, objectives, and bibliography. No In synthesis, academic or research integrity
human protocol is required. References to study as a terminology or concept plays little part in the
samples, human subjects, and data collection actual institutional culture, or at least is not
procedures, when indicated, usually are located configured as a formal organizational principle in
in a section on “methodology.” the university culture. This is not to say that
Research projects involving human subjects academic integrity is not present in many of the
must have the approval of the Committee on pedagogical and academic actions of students
Ethics in Research only for professors from the and faculty, nor in the daily practices of this
Biomedical Center. This Committee was institutional culture. Nevertheless, the fact that
registered in March of 1997. No communication academic/scientific ethics or research integrity
from this committee to other academic centers are not explicitly registered in formal university
has been documented by the institution. The institutional norms considerably complicates the
potential institutional role of this committee institutional capacity to develop scientific
could be to distribute and discuss the present integrity and deal with ethical problems of any
regulations, which affect other areas of nature.
knowledge.
The lack of information on the necessity for Conclusions
compliance with existing regulatory standards for These results confirm the necessity for urgent
human protocol or the absence of academic/ institutional action to establish normative
administrative requirements for recognizing the standards that promote a responsible research
ethical consideration of data collection with environment and a critical consciousness of the
human subjects are seen as substantial obstacles need for training/research in scientific integrity in
for promoting research integrity. However, the all areas of knowledge. However, the
implications for dealing with possible advancement of academic/scientific ethics
misconduct are the most serious. depends upon a critical analysis of present
The first dilemma is the extreme negligence research practices and the recognition of the
with which most universities treat their internal protection of human subjects as one component
problems of human communication and of research integrity inherently connected to the
academic relationships among faculty and ethical production of knowledge.
students, with no viable procedures or Institutional research is needed to identify
mechanisms to identify, solve, or prevent such academic areas with accessibility for a new
problems. In the case of the public Federal approach to teaching research integrity as well a
universities, professors and university current researchers’ concerns with research
functionaries are classified, by law, as federal ethics. Institutional support for such curriculum
public servants, subject to Federal legislation. reform is vital, but must occur with a greater
The legislation is basically a disciplinary regime strategy to set university goals for excellence in
where duties and obligations are specified. research with human subjects and to reform
Denouncements of irregularity/misconduct are regulations that are obsolete and ineffective in
treated administratively in an process that can dealing with problems of academic/scientific
consume a year or more. integrity.
These laws as well as the university statues Caution is necessary to avoid
and internal regulations date from the years of “overdeveloped” procedures that do more to
the military dictatorship in Brazil, seldom having serve the rule makers than to protect the victims
been reformed to establish a less authoritarian of unethical research practices. Perhaps, instead
academic administrative structure. These of taking the long road and merely reproducing
instruments refer to problems with faculty or regulations and administrative procedures for
student behavior in terms of order and discipline, projects review, or awaiting federal legislation,
keywords common to public policy of the local universities such as the UFES should
military government. Academic problems consider the middle road, one which is not a
102
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Doxsey, Promoting Scientific Integrity
short cut or dodges vital issues, but one which
stimulates a process that provides access to
information, provides debate about research
integrity, and acknowledges institutional needs
for guidelines to avoid scientific misconduct and
to safeguard human subjects, particularly those
subjects in situations of cultural or social risk.

Bibliography
1. Bontempo de Freitas C, Oliviera Lôbo M. CONEP faz
balanço positivo [National Commission of Ethics in
Research – CONEP makes a positive evaluation].
Cadernos de Ética em Pesquisa. 2000;4;4-7.
2. A ética e a bioética em livros [Ethics and bioethics in
books] [Special article]. Cadernos de Ética em Pesquisa.
2000;5;17-19.
3. A ética nas ciências humanas [Ethics in the human
sciences] [interview]. Cadernos de Ética em Pesquisa.
2000;4;14-7.
4. Sardenberg T, Müller SS, Pereira HR, Oliveira RA,
Hossne WS. Estudo mapeia ética em revistas [Study
maps ethics in scientific journals] Cadernos de Ética em
Pesquisa. 2000;5;11-14.
5. Doxsey JR. Ética academica e cientifica numa
perspectiva psicossociológica: a escuta ativa do ouvidor
[Academic and scientific ethics in a psycho-sociological
perspective: the active listening of an ombudsman]. In:
Pinto Lyra R, organizer. A ouvidoria na esfera pública
brasileira [The ombudsman in the Brazilian public
domain]. João Pessoa (PB): Ed. Universitária, 2000. p.
143-57.

103
Ethical Research Practice with Human Participants: Problems, Procedures,
and Beliefs of Funded Researchers
Elana Newman, Department of Psychology, University of Tulsa, USA

Victoria McCoy, Department of Psychology, University of Tulsa, USA

Anthony Rhodes, Department of Psychology, University of Tulsa, USA

Keywords: Cognition, Confidentiality, Ethics, Health, Informed consent, Mental health.

Although Federal and local guidelines provide general advice as to inform researchers regarding
ethical practice (1 - 3), little information is available regarding how researchers carry out such ethical
procedures. Despite the use of Institutional Review Boards (IRBs) to monitor ethical practice, there
is great variability in how these boards operate and what types of policies are deemed acceptable (4).
Similarly, it appears that psychopathology researchers greatly differ in their practices on how to
assess and handle participant distress or injury (5 - 7). In some specialty areas, such as depression,
there is preliminary evidence that most researchers routinely give referrals (8). Nevertheless, the
range of practice is not known.
The need to document how different biomedical researchers implement ethical research policies
is important in order to generate and develop viable and informed research policy. For example, it is
helpful to understand how researchers recruit participants, train staff, obtain informed consent, and
debrief participants (9). Furthermore, specific policies about response and compensation with regard
to responding to participants’ distress, worsening of conditions, confidentiality issues, informed
consent, and other ethical dilemmas across different groups of human research participants is also
needed. Sharing such information among researchers from different disciplines, who use different
methodologies and research samples, can help to identify the range of options and the need for
training initiatives. Finally as technology makes research more global, local community standards of
practice may no longer be adequate to understand good research practice (10). To compound this
issue, distinctions between research and clinical work and research and organizational consulting are
blurring with the trends in program evaluation. Finally, advances in science have made human
experimentation itself more complex. Hence there is a need to share information and understand the
range of ethical practice in the field so we are better able to respond to these challenges and equipped
to create policy in the future.
Currently it is unknown how often research-related injuries and problems occur in the course of
routine research protocols. Although flagrant violations are reported or receive media attention, there
has been no attempt to quantify the prevalence of such problems in routine practice (11). In order to
understand participants’ responses it is also important to ascertain the actual prevalence rates of
research-related costs and injury across a wide range of samples to determine what groups need
Corresponding author: Elana Newman, Ph.D., Department of Psychology, University of Tulsa, 600 South College Avenue ,
Tulsa, OK 74104, 918-631-2248 (voice), 918-631-2833 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
additional safeguards. These risks must be researchers who studied humans with
quantified to include both minor costs (abrasions, schizophrenia (n = 264), cardiovascular disease
emotional distress) and major costs (death, (n = 1472), major affective disorder (n = 899),
disability, and needed hospitalization). and traumatic stress (n = 564) were identified
Identification of the subgroups at greatest risk for from relevant NIH institutes using the
research related harm could help inform policy Community of Science National Institute of
(12). Health database of funded grants (http://
Finally the expertise of researchers and cos.gdb.org/best/fedfund/nih-select/inst.list.html)
opinions need to be shared. As documented, and the Veterans Administration Medical Center
opinions and assumptions about possible risks grant database (http://www.va.gov/research/
and benefits of research participation shape research.html). These groups were chosen to
ethical appraisals of research (13 - 17). represent medically and psychiatric samples that
Documenting experienced scientists’ opinions are hypothesized to be at greater risk for
and attitudes toward IRBs and research risk, can research-related injuries. In addition, we
help establish a clearer understanding of the identified a pool of 485 federally funded
values that may shape research and research investigators who study cognition in non-patient
policy. samples to represent a group hypothesized to be a
The goal of the current study is to delineate relatively lower risk for research-related
the rates and types of potential research-related research.
injuries as well as the range of ethical practices Relevant grant proposals were identified by
and beliefs. This is important since several conducting a search of all proposals that had
studies document the range of ethical research titles which contained a relevant key word. For
practice, but none of them actually assess the example for studies on depression, depression
prevalence and types of risks (8). needed to be in the title. For traumatic stress
First, it was hypothesized that there is studies, PTSD, trauma or stress needed to be in
considerable variability of research policies and the title. A detailed listing of key words and the
procedures both within and across types of systematic manner in which certain protocols
research and sample characteristics with those were eliminated is available from the first author.
researchers working with psychiatric illness Studies that crossed topic domains, used minors,
being more protective than researchers in other used animals, or were post-mortum human
areas. Policies and procedures were defined as studies were eliminated from the pool of studies.
(a) level informed consent policy, (b) emergency All treatment studies were eliminated, since they
policies, (c) determination of research- related have unique risks and benefits that were not
risk, (d) debriefing procedures, (e) use of assessed in this study. All projects that were
referrals, and (f) follow-up procedures. funded as multi-site collaborative studies were
Second, it was hypothesized that the research also eliminated since it was assumed the ethical
risks experienced by psychiatric health groups considerations might vary across site.
will be significantly greater than those Ultimately, 69 funded researchers who study
experienced by the medical physical health cognition, 79 who study schizophrenia, 61 who
group. In addition, it was hypothesized that study lung-cardiovascular disease, 56 who study
researchers who studied psychiatric and medical affective disorders, and 49 who study violence/
samples were expected to report significantly PTSD were contacted.
greater rate of research risks than the non- A cover letter, 7 page survey form1 , and
psychiatric or medical samples. Research risk return envelope were sent to 314 researchers. A
was defined as (a) Incidence of confidentiality reminder card was sent one month later to all
violations for suicide, homicide, and abuse responders and non-responders. The survey
status; (b) Incidence of participants’ condition began with general information about the
worsening; and (c) Incidence of complaints and respondent’s demographics, and research and
or suits filed against researcher or institution. clinical experience. The researcher was asked to
complete the questionnaire in regard to the most
Method recent funded grant. Questions pertained to the
We generated a list of 3,684 investigators who setting, sample, type of research, number of
received federal funding for research projects sessions, participant characteristics, staff/training
pertaining to four at-risk groups. Specifically, and supervision. Then questions about informed
106
–––––––––––––––––––––––––––––––––––––––––––––––– Newman, et al., Ethical Research Practice with Human Participants
consent, confidentiality issues encountered, Although 85% reported no need to investigate if
participants’ reactions, emergency policies, and the identified participant could legally provide
injuries were attached. consent, the remaining 15% reported a need
ranging from once (7%) to eighty-five times
Results (1%).
A total of 101 surveys were returned yielding a With respect to informed consent, 53% of
32% response rate. Eleven surveys were dropped these researchers indicated that there were
from the analysis because they were post-mortem instances in which the confidentiality of the
studies (n = 4), used minors exclusively (n = 1), research participant might be broken. As
focused on substance abuse, HIV, or personality predicted, this policy differed by type of sample
disorders (n = 4), animal studies (n = 1) or group [x2 (2, n = 85) = 10.75 p =<.05], with 66%
couldn’t be classified into the groups based on of those who worked with mental health groups,
the responses (n = 1). Of the 9 researchers who 55% of those who worked with physical health
participated, 52.2% studied mental health (PTSD groups, and 21% of those who studied cognition
n = 12, schizophrenia n = 16, major affective stating instances in which the research team
disorders = 19), 24.4% studied cardiac or health would consider breaking the confidentiality of
problems and 23.3% studied “normal” cognition. the research record. Among the group who
informed participants about confidentiality
Participants issues, 55% reported communicating this in
The 90 principal investigators were comprised of specific rather than general terms.
primarily Ph.D. trained researchers (73%) and Emergency Policy. Seventy-eight percent
M.D.s (19%). There were more males (63%) (n = 61) of the researchers endorsed having a
than females (37%) represented, and the majority protocol in place a priori to respond to
of respondents were Caucasian (94%). The emergencies. The groups significantly differed
respondents’ experience with research ranged in this policy [x2(2, n =78) =32.15, p <.05] such
from 2 to 49 years and had received a mean of that 95% of mental health researchers, 90% of
2.8 (SD = 1.8) federally funded grants in the 5 physical health researchers, and 28% of cognitive
years prior to the study. The group of researchers researchers reported such emergency policies in
reported a mean of 70 peer-reviewed place. Among the 47 who provided open ended
publications, a median of 44 and a mode of 150. descriptions of these policies, 15 described use of
Only 20% reported completing a course in emergency on-call personnel, 8 cited they had
research ethics during advanced training. “written policies,” 6 used standard local
Despite this lack of formal training, 73% felt that protocols, 6 cited immediately contacting the
they kept current with ethical issues and 50% felt project director or principal investigator, 5
they kept current with legal issues in research. trained staff in Cardio Pulmonary Resuscita tion
Only 6% and 22% felt they were not current (CPR), and 3 discussed continuous monitoring
regarding ethical and legal research issues, during research. The remaining four described
respectively. emergency medication, medical response plan in
lab and for evacuation, methods for handling
Research Procedures high blood pressure, and one general training
Informed Consent Policy. With respect to how to respond to a variety of situations.
informed consent, the majority of the sample Determination of Research-Related Risk.
(97%) provided written informed consent and Seventy-eight percent (n = 62) of the researchers
48% endorsed using methods to assess sampled reported keeping records regarding the
participants’ comprehension of the consent form. “frequency to which individuals experienced
Of the 39 respondents who provided open ended negative and noticeable reactions.” Mental
descriptions of these methods, 25 asked health researchers reported significant greater
participants if they had questions, 3 had the documentation than health or cognitive
interviewer certify person heard and understood, researchers [x2 (2, n = 81) = 19.79, p < .05] such
3 used independent monitors, 2 relied on other that 88% of mental health researchers, 79% of
indicators (fluency, literacy, neurological status), physical health researchers, and 52% of cognitive
1 used family consent, 1 used structured consent, researchers kept such records.
2 asked the respondent to repeat questions, and 2 Debriefing Procedures. Sixty-four percent
relied on signature to indicate comprehension. (n = 57) of the researchers conducted debriefings
107
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Ranki ng
Factors Least important Important Fairly Important Most Important
Manipulation check 24 (63%) 5 (13%) 8 (21%) 1 (3%)
Educate participants 1 (2%) 18 (33%) 7 (13%) 28 (52%)
Check on participant 7 (14%) 12 (24%) 10 (20%) 21 (42%)
Express gratitude 6 (11%) 9 (16%) 26 (46%) 15 (27%)

Table 1. Number (and percentage) of participants ranking relative importance of 4 factors in planning debriefing.

after the research protocol. In fact, 70% of blood pressure.”

mental health professionals, 42% of health Given this practice, the number of referrals
researchers, and 71% of cognitive researchers for non-emergencies ranged from 0 to 90 (mean
used such debriefings [x2 (2, n = 80) = 5.06, = 4.76, s.d. =13.02; mode =0). The mean
p =.08]. The majority (80%) of these number of referrals for the mental health, health
debriefings were conducted orally, although 6% and cognitive research teams were 8.56 (S.D. =
were conducted in writing, with 14% conducted 17.83), 2.29 (S. D. = 4.10) and .40 (S.D. =1.05)
in both formats; there was no statistically respectively, but these differences did not meet
significant difference among the groups criteria for statistical significance [F (2, 65) =
regarding format [x2 (4, n = 51) = 4.48, p = .34]. 2.9, p = .062].
The majority of these debriefings were done in With respect to actual practice regarding
individual sessions (88%) rather than group referral for immediate hospitalization, 6
(4%), varied (6%) or family formats (2%); this researchers recommended immediate referral for
did not vary significantly among groups format a condition or concern, (with two researchers
[x2 (6, n = 51) = 9.05, p = .17]. As can be seen on recommending it once, and the rest experiencing
Table 1, investigators felt debriefings were most it twice, three times, four times and 10 times). It
important for educating participants and checking is unknown if these referrals were based on
on participants. It is interesting to note that research-related injuries, or other conditions
manipulation checks were deemed least uncovered during the protocol.
important. Follow-up procedures. Fifty-four percent
Use of Referrals. Forty-one researchers (n = 41) of the researchers reported follow-up
(46% of the sample) responded to the item about efforts to determine if participants experienced a
referral policy. Among those who responded, worsening of condition. These efforts
20% reported providing referrals to all significantly differed across groups [x2 (2, n = 76)
participants, 12% to those participants who = 14.35, p <.01] such that 67% of mental health
indicated interest, 17% to only those in distress, researchers, 55% of health researchers, and 8%
42% to those either interested or distressed, and of cognitive researchers used such methods. In
10% in “other” circumstances. Three researchers terms of actual numbers, 24 researchers reported
described such other circumstances as “offered to conducting a follow-up at least once to check on
all deemed appropriate, but given to those a participant.
interested;” “two found to have physical
disorders,” and “all those screened with high

Never Infrequently Sometimes Regularly Always

Suicidality 58 (64%) 20 (24%) 4 (4%) 2 (2%) 1 (1%)
Homicide 76 (91%) 5 (6%) 2 (2%) 1 (1%)
Child abuse 72 (85%) 9 (11%) 2 (2%) 2 (2%)
Elder abuse 78 (94%) 4 (5%) 1 (1%)
Abuse of the disabled 78 (94%) 4 (5%) 1 (1%)
HIV status 64 (77%) 9 (11%) 8 (10%) 2 (2%)
Substance abuse 49 (59%) 10 (12%) 14 (17%) 9 (11%) 1 (1%)
Criminality 68 (83%) 9 (11%) 1(1%) 3 (4%) 1 (1%)
Violence toward partner 67 (80%) 11 (13%) 2 (2%) 3 (4%) 1 (1%)
Other 50 (94%) 3 (6%)
Table 2. Number and (Percentage) of researchers who faced confidentiality issues.

108
–––––––––––––––––––––––––––––––––––––––––––––––– Newman, et al., Ethical Research Practice with Human Participants

Research Risks of research conducted [x2 (2, n = 73) = .42, p =

Incidence of confidentiality violations. The .81]
research staff occasionally faced confidentiality Incidence of complaints filed against a
dilemmas as shown in Table 2, with substance researcher or institution. In this sample, 18%
abuse being the most frequently encountered reported infrequent complaints about research
issue. However, only 8 researchers actually staff’s conduct. Two percent (n =2) reported
broke confidentiality. Of these 8, 6 studied complaints filed against the institution however
mental health (n = 3 mood disorders, n = 2 none resulted in legal proceedings. On the other
schizophrenia, n =1 PTSD), 1 studied normal hand, 77% of researchers reported that
cognition, and 1 studied health conditions. participants thanked them, with 33% reporting
Among those researchers who described the this occurring sometimes, and 12% reporting this
specific circumstances, two reported needing to as a regular occurrence.
hospitalize at least one participant against his/her
will, three reported having to file at least one Discussion
report to the authorities, and two reported In this preliminary study, 90 federally funded
needing to warn at least one person in danger. researchers who work with human participants
Incidence of participants condition responded to a survey about ethical research
worsening. During the protocol a range of practice. There seems to be a great variation in
emotional and physical experiences were ethical practice among distinguished researchers,
encountered (See Table 3); clearly crying although all these research participants were
appeared most often. Although it was rare that a sensitive to research-related ethical dilemmas.
participant became medically compromised, it
did occur. Twelve researchers (13%) reported at Policies
least one research-related injury. Two researchers There is a great deal of variation in research
reported that at least one participant had a policy implementation. Although nearly all use
research-related infection. Five researchers written informed consent, researchers varied in
reported at least one case of temporary disability, the detail that they provide participants about the
and none reported research-related death. It limits of confidentiality. Although the majority
should be noted that only 53% of researchers of researchers developed emergency policies and
reported knowing how many participants debriefing procedures, the nature of these
experienced an immediate worsening of procedures also varied. Although often required,
condition (research related injuries) after 32% did not keep records of participants’
completing the research protocol; Knowledge of negative and noticeable reactions.
research-related injuries was not related to type Approximately half the researchers reported

Never Infrequenty Sometimes Regularly Always

Cried 35 (42%) 24 (29%) 16 (19%) 7 (8%) 1 (1%)

Became hostile or 33 (43%) 35 (42%) 13 (16%) 3 (2%) 0

angry
Experienced Panic 59 (71%) 17 (21%) 6 (7%) 1 (1%) 0
Attacks
Expressed extreme 55 (66%) 16 (20%) 8 (9%) 4 (5%) 0
fear
Reported feeling 51 (62%) 18 (22%) 12 (15%) 1 (1%) 0
spacey
Became medically 66 (81%) 14 (17%) 2 (2%) 0 0
compromised
Threatened the 71 (87%) 10 (12%) 1 (1%) 0 0
research staff
Other 33 (86%) 2 (5%) 1 (3%) 1 (3%) 1 (3%)

Table 3. Number and percentage of researchers who encountered participants’ emotional or physical response to research.
109
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
using follow-up methods to check on implementation is not examined. Hence it is not
participants’ condition. However, less than half known if this diversity suggests unsuccessful or
the sample responded to the item regarding the successful flexibility of methods in responding to
use of referrals and those that did respond the needs of human participants.
indicated a range of practices with respect to Although the participation rate precludes
referring to other agencies. As anticipated, generalizing to all researchers, these preliminary
researchers working with psychiatric illness results provide information that can be useful in
being more protective and explicit about policies designing training and compliance policy. In
for emergencies, risk documentation, and follow- particular, the diversity of responses suggests the
up procedures but not for debriefing. need for cross-training across subspecialties to
share perspectives. Individuals with risk factors
Risks may not only present for studies of health and
With respect to research risk, a minority of mental health problems, so it can be helpful to
researchers reported having to deal with share approaches across specialties. For example,
confidentiality issues, worsening of conditions, although the majority of research-injuries were
and complaints from participants. However, identified among those mental heath studies, they
emotional and physical symptoms were were not exclusively there. Furthermore it is
encountered. In particular, 58% (n = 48) unclear, given the lack of documentation and
experienced crying, and 12 researchers (13%) investigation, if this reflects better preparedness
reported temporary research-related injuries. of mental heath researchers or greater risk in
Given that several of these studies were about these studies. Future studies may be able to
health conditions, it is difficult to evaluate if better examine this by ongoing quality control
these reactions were elicited by research (19).
participation, or were symptoms that individuals
experienced irrespective of research Acknowledgements
participation. These reactions need to be The research for this paper was supported by a
examined in future studies in the context of University of Tulsa Summer Faculty
baseline functioning of individuals to further Development Grant. A portion of this paper was
understand if they meet the requirements of presented at Research Conference on Research
minimal risk. Nonetheless, the data are Integrity, Office of Research Integrity, Bethesda,
consistent with claims that the physical hazards MD, November 2000.
of being a research participant are minimal even
among medical procedures (18). Although, these Notes
risks appear minimal, they might be an 1. A copy of the survey is available from the first author.
underestimate given that about half the
researchers did not document or know the Bibliography
number of participants whose condition 1. American Psychological Association. Ethical principles
worsened. of psychologists. Am Psychol 1981; 36: 633-8.
Finally, very few researchers received formal 2. Protection of Human Subjects. 45 C. F. R. Subtitle A,
training in research ethics although the majority pt. 46. (2001).
3. National Commission for the Protection of Human
were confident that they were up to date in
Subjects of Biomedical and Behavioral Research. The
ethics, and half felt prepared for legal challenges. Belmont Report: Ethical principles and guidelines for
Given that researchers thought highly of their the protection of human subjects of research (DHEW
respective IRBs, continuing education may be Publication No OS 78-0012). Washington DC:
best implemented through local IRBs. Department of Heath, Education, and Welfare; 1978.
There are several limitations to this study. 4. Goldman J, Katz MD. Inconsistency and institutional
First sample bias and demand characteristics may review boards. JAMA 1982; 248(2): 197-202.
have affected the generalizability of these results. 5. Barton JM, Macmillan MS, Sawyer L. The
Although the extensive comments written on compensation of patients injured in clinical trials. J
Med Ethics 1995; 21(3): 166-9.
those returned surveys suggest that researchers
6. Burbach DJ, Farha JG, Thorpe, JS. Assessing depression
were interested in sharing their experiences, in community samples of children using self-report
sample bias may have affected the results. inventories: Ethical considerations. J Abnorm Ch
Second, while this study reveals a diversity of Psychol 1986; 14(4): 579-89.
ethical practices, the quality of ethical
110
–––––––––––––––––––––––––––––––––––––––––––––––– Newman, et al., Ethical Research Practice with Human Participants
7. McCrady BS, Bux DA. Ethical issues in informed
consent with substance abusers. J Consult Clin Psychol
1999; 67: 186-93.
8. Stanton AL, New MJ. Ethical responsibilities to
depressed research participants. Prof Psychol 1988; 19:
279-85.
9. Silverman HJ. Ethical considerations of ensuring an
informed and autonomous consent in research involving
critically ill patients. Am J Respir Crit Care Med 1996;
154(3): 582-6.
10. Edgar H, Rothman DJ. The institutional review board
and beyond: Future challenges to the ethics of human
experimentation. Milbank Q 1995; 73: 489-96
11. Hilts PJ. Safety Concerns Halt Oklahoma Research.
The New York Times 2000 Jul 11; Sect. A:12.
12. Newman E, Kaloupek DG, Keane TM, Folstein S.
Ethical decision making among trauma researchers. In
Kantor GK, Jasinski J, editors. Out of the darkness:
contemporary perspectives on family violence.
Newbury Park (CA): Sage Press; 1997; 271-81.
13. Ceci, SJ, Peters D, Plotkin J. Human subjects review:
personal values, and the regulation of social science
research. Am Psychol 1985; 40(9): 994-1002.
14. Oldham JM, Haimowitz S, Delano SJ. Protection of
persons with mental disorders from research risk. Arch
Gen Psychiatry 1999; 56: 688-93.
15. Scarr S. Race and gender as psychological variables:
Social and ethical issues. Am Psychol 1988; 43: 56-9.
16. Sieber JE. Issues presented by mandatory reporting
requirements of researchers of child abuse and neglect.
Ethics Behav 1994; 4: 1-22.
17. Stanley B, Sieber JE, Melton GB. Empirical studies of
ethical issues in research. A research agenda. Am
Psychol 1987; 42(7): 735-41.
18. Levine M. Ethics and regulation of clinical research.
Baltimore: Urban & Schwarsenberg; 1981.
19. Newman E, Willard T, Sinclair R, Kaloupek D
(forthcoming). The costs and benefits of research from
the participants’view: the path to empirically informed
research practice. Accountability Resear 2001; 8: 27-
47.

111
Balancing Risks and Benefits of Deception in Assessing Genetic Screening*
Dennis L. Thombs, Department of Adult Counseling, Health and Vocational Education, Kent
State University, USA
Colleen A. Mahoney, Mahoney Consulting Group, Kent, Ohio USA
R. Scott Olds, Health Education and Promotion, Kent State University, USA

Keywords: Alcoholism, Deception, Genetic screening

The Human Genome Project is a massive international research program designed to map the human
genome sequence(1). The fundamental purpose of the program is to spur a transition to DNA
sequence-based biology and biomedical science(2). In addition to revolutionizing medical
diagnostics and therapy, the Human Genome Project will create new challenges in a variety of fields
including law, medical ethics, public health, and health services administration(3). The anticipation of
these changes does not represent a distant concern. A “working draft” of the entire human sequence is
expected by the end of 2001(2).
Against the backdrop of the Human Genome Project, this article critically examines the use of
intentional deception to assess (and anticipate) the utilization of genetic screening for alcoholism
susceptibility. For some time, the manipulation of study participants by deception has been controver-
sial in experimental social psychology(4). This controversy has emerged in health behavior research
as a consequence of the remarkable progress made by the Human Genome Project. Little is known
about the public’s interest and utilization of clinical genetic testing(5). In the specific area of predic-
tive genetic screening, a deception paradigm (described below) has been found useful for assessing
utilization. This paradigm helps estimate utilization when such tools are on the horizon, but not yet
available to the consumer. Intentional deception appears to be necessary because “hypothetical
testing,”(6, 7) honestly described to research subjects as available “sometime in the future,” generates
inflated interest compared to testing described as “currently available”(8, 9).
In an editorial that appeared in the Journal of American College Health,“Hard Questions About
Research Procedures: The Search for Authenticity”(10), Dr. Richard Keeling objected to the use of
deception in a quasi-experimental study conducted by the authors. The report of this investigation
appears in the same issue of that publication “Application of a Bogus Testing Procedure to Determine
College Students’ Utilization of Genetic Screening for Alcoholism”(11). Interested readers may turn
to that article for a full description of the study methods, including the fabricated story concocted to
test student interest in genetic screening for alcoholism susceptibility.
Dr. Keeling’s editorial is an example of a conservative, but perhaps increasingly common position

* Reprinted with permission from the Am J Health Behav 2001; 25(2): 100-105. R. Scott Olds was invited to present this
paper at the RRI conference under the title: Responsible Research Conduct that Balances Risks and Benefits of Deception in
Assessing Genetic Screening Utilization.
Corresponding author: R. Scott Olds, H.S.D., Health Education and Promotion, 316 White Hall, Kent State University,
Kent, OH 44242-0001, 330-672-0679 (voice), 330-672-3063 (fax) , [email protected].
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
on human subjects protection that exaggerates risks and benefits has tilted toward the former in
risk to study participants and discourages poten- recent years. If so, does this shift represent
tially valuable inquiry. The conservative position increased concern for human subjects? An
is based on the following beliefs: 1) deception is iconoclastic interpretation is that the conservative
inherently harmful; and 2) deception research is analysis of risk has been motivated by fears of
not carried out under realistic conditions and lawsuits and a desire to protect the university
therefore is not of value. The authors believe from legal action. In addition, doubts about the
their views are based on an ethic of measured and quality and usefulness of behavioral science
reflective discourse, instead of a “knee-jerk” research in general, may be in operation in some
response fashioned to serve a particular ideology. quarters which only further discourages full
According to Aronson and colleagues (4), consideration of the potential benefits of such
when considering the use of deception in re- work.
search, investigators must weigh the psychologi- No data were collected in this study to
cal discomfort participants may experience support the claim that the students were not
against the value of the study. There is no single harmed by the deception. However, it should be
set of rules that can be applied to resolve this noted that the empirical literature does not
dilemma, and reasonable professionals will arrive support the view that research using deception is
at different judgments in this difficult analysis. any more harmful than non-deception research
To determine college student interest in genetic (4). One review of the literature concluded that it
screening for alcoholism susceptibility, it was was rare for participants to feel that they had
reasonable to expose them to what was believed been harmed by intentional deception (12).
to be modest psychological and social risks. The Though empirical studies on the effects of
Institutional Review Board at Kent State Univer- deception are few, those that have been con-
sity concurred, and with certain stipulations gave ducted generally have found that participants
approval to conduct the study. report greater enjoyment from having partici-
The subjects in this study were deceived pated in a deception experiment than in a
about the availability of a genetic screening test. nondeception experiment (13). This is probably
For up to seven days, 181 students thought they due to deception studies being less boring (4). To
could schedule a predictive screening test for address these concerns, in the future, investiga-
alcoholism that does not yet exist. The authors tors should follow up with participants to deter-
did not believe that this lie harmed the students mine their reactions to research deceptions.
in any substantial way. In broad-brush comments, It is noted that the source of discomfort in
Dr. Keeling (10; see page 101 of his editorial) deception research is not only learning later that
claims that today’s college students are often one has been deceived, but equally, if not more
exploited by society and that any challenge to important is that the person often learns some-
their “search for authenticity” poses an unaccept- thing painful about themselves or others (14).
able risk to their mental health and/or future Again, data were not collected to support this
social functioning. It seems that this view is not hypothesis, but it is strongly suspected that
unusual in academia today. Such a position among those students who were uncomfortable in
represents “politically correct” discourse that this study, the primary source of their discomfort
exaggerates the risks of deception in this study was their current drinking behavior. As noted, the
and casts a broad net of condemnation over all sample was over-represented by heavy drinking
uses of deception in research. Clearly, humans students. Participation in the study required them
have been mistreated in research that employed to reflect on their own alcohol use as well as that
deception (e.g., the Tuskegee Syphilis Study), but of their family members. Indeed, it was sensed
distinctions can and should be made in its by the authors that some students were uncom-
application. fortable while responding to the questionnaire
In this era of heightened concern about and watching the presentation. In other words,
compliance with Federal regulations on research the discomfort that some experienced appeared to
involving human subjects, “minimal risks” in occur before the debriefing, rather than after it
behavioral science research have sometimes been (when they learned they had been deceived).
subtlety redefined as “unacceptable risks.” The Some students actually appeared amused during
authors have no data to support or dispute such the debriefings.
speculation, but wonder whether the balancing of The level of discomfort experienced by
114
––––––––––––––––––––––––––––––––––––––––––––––– Thombs, et. al, Deception in Assessing Genetic Screening Utilization
students was probably comparable to being asked participation at any time without penalty.
to participate in an anonymous self-report survey Dr. Keeling was accurate in describing that over
of alcohol use, and probably no greater than the next seven days, students were not given
sitting in routine lectures and discussions in counsel or additional information about the test.
health education courses that deal with any In this respect, the procedure was not as realistic
number of sensitive issues. The discomfort that as future testing probably will be, but neither was
some may have experienced was not considered it as unrealistic as described by Dr. Keeling in his
to be detrimental or bad. Good health education editorial. It is acknowledged that in the future,
“shakes up” students by confronting biased people may contemplate the testing decision for
perceptions of risk and challenging existing extended periods of time, perhaps even many
social norms. It also is consistent with the years. Obviously, this study does not address
traditional view of higher education, which is to readiness to seek testing over extended time
challenge conventional thinking and behavior intervals, but it does provide marketing
and to engage students in debate about controver- information about what to expect if promotion of
sial issues. genetic screening for alcoholism susceptibility
Dr. Keeling (10) also was critical of the among high-risk drinkers becomes a public
contention that the study conditions were “realis- health goal.
tic.” The authors agree with his observation that The preliminary findings from this study
if (or when) genetic testing for alcoholism suggest that among college students, there may
susceptibility becomes available, protocols very be little enthusiasm for seeking alcoholism
likely will require extensive patient counseling screening if (or when) it becomes available.
before and after the procedure. So by this bench- Certainly this issue deserves further investiga-
mark, the study’s procedure was not realistic. The tion. The authors believe the health promotion
authors should have been more precise by stating profession has an obligation and responsibility to
that “our method was more realistic than using a conduct research that anticipates and informs the
procedure that described screening as a future development of sound public health policy. If
possibility.” However, at the same time, introduc- future public health policy supports genetic
ing extensive patient counseling into the study testing for alcoholism susceptibility, ethical
procedure would have required us to employ a questions need to be raised by the professions
far greater level of deception. Such a research concerned with public health. This study is part
design would be considered unethical by virtu- of the foundation needed to address these ques-
ally all professionals and would justify tions.
Dr. Keeling’s response. This study protocol, These debates are important and healthy, but
however, does not. they are not easy. The issues surrounding genetic
As the study was carried out, participants testing are complex. Billions of dollars are being
were deceived for no more than seven days. They spent on genome research for the purpose of
were debriefed and offered the opportunity to developing effective technologies to treat and
withdraw their data without penalty. In his prevent disease. Yet, relatively little attention is
editorial, Dr. Keeling (10) stated, being given to the behavioral, social, and health
. . . Having watched a computer-generated service implications of this technology. There is a
presentation (for 7 minutes) and heard a brief need to better understand the utilization of
explanation of the study itself, they were then predictive screening for a variety of disorders,
required to state their intentions about being including alcoholism. This study should stimu-
tested immediately. There was little time for late discussion among health promotion profes-
them to ponder the issues and submit a formal
request to be tested. . .(p. 100).
sionals about these aspects of genetic testing.
This description of the study’s methods is not
accurate. Careful reading of the methods clearly Bibliography
1. National Human Genome Research Institute. The
stated that students were told they did not have to
Human Genome Project. www.nhgri.nih.gov, 2000;
make a decision immediately after the April, 14.
presentation. A questionnaire item allowed them 2. Collins FS, Patrinos A, Jordan E, Chakravarti A,
to respond I am uncertain about whether or not Gesteland R, et al. New goals for the U.S. Human
to be tested (see p.106 of our article)(11). Genome Project: 1998-2003. Science
Further, their participation was always voluntary 1998;282:682-689.
and invitational. They were able to cease
115
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
3. Khoury MJ and Genetics Work Group. From genes to
public health: The applications of genetic technology in
disease prevention. Am J Pub Hlth 1996;86:1717-1722.
4. Aronson E, Ellsworth PC, Carlsmith JM, Gonzales MH.
Methods of Social Psychological R esearch (Second
Edition). New York: McGraw-Hill; 1990.
5. Croyle RT, Lerman C. Interest in genetic testing for
colon cancer susceptibility: Cognitive and emotional
correlates. Prev Med 1995;22:284-292.
6. Kessler S, Field T, Worth L, Mosbarger H. Attitudes of
persons at risk for Huntington’s disease toward
predictive testing. Am J Hum Genet 1987;26:259-270.
7. Williamson R, Allison MED, Bentley TJ, Lim SMC,
Watson E, et al. Community attitudes to cystic fibrosis
carrier testing in England: A pilot study. Prenat Diag
1989;9:727-734.
8. Craufurd D, Dodge A, Kerzin-Storrar L, Harris R.
Uptake of presymptomatic predictive testing for
Huntington’s disease. Lancet 1989;2:603-605.
9. Tambor ES, Bernhardt BA, Chase GA, Fadden RR,
Geller G, et al. Offering cystic fibrosis carrier screening
to an HMO population: Factors associated with
utilization. Am J Hum Gen 1994;55:626-637.
10. Keeling RP. Hard questions about research
procedures:The search for authenticity. J Am Coll
Health 1998;47:99-102.
11. Thombs DL, Mahoney CA, Olds, RS. Application of a
bogus testing procedure to determine college students’
utilization of genetic screening for alcoholism. J Am
Coll Health 1998;47:103-112.
12. Baumrind D. Research using intentional deception:
Ethical issues revisited. Am Psychol 1985;40:165-174.
13. Christensen L. Deception in psychological research:
When is it justified? Personality and Social Psychology
Bulletin 1988;14:664-675.
14. Smith SS, Richardson D. Amelioration of deception and
harm in psychological research: The importance of
debriefing. Journal of Personality and Social
Psychology 1983;44:1075-1082.

116
Research Integrity and the Direct Involvement of Persons with Disabilities
Nancy B. Robinson, Communication, Arts and Sciences Department, California State
University, Chico, California, USA
Kathleen C. Sadao, Curriculum and Instruction Department, University of the Pacific,
Stockton, California, USA

Keywords: Responsible conduct of research, Preservice training, Problem-based learning, Persons with
disabilities, Affective learning

Teaching students in the health, human service and education professions to be responsible in their
interactions with persons with disabilities, as service providers and researchers, poses unique
challenges to educators to move beyond imparting knowledge to impacting attitudes, values and
ethics. Recent emphasis on outcomes of professional education programs most frequently focuses on
indices of cognitive achievement and performance of specific skills or competencies. Measures of
affective learning, or student attitudes and values toward the persons they serve, are less frequent and
more difficult to document. Universities need to educate professionals who are capable of the
responsible conduct of research. Pre-service education models are shifting from a traditional didactic
approach to the use of case studies and problem solving, in an effort to influence affective learning
and the application of knowledge and skills in real-life simulations. Studies of effective teaching
methods to prepare professionals in the area of responsible conduct of research with human subjects
are clearly needed. Person-focused learning approaches developed from interactive teaching models,
used increasingly in pre-service education in disability services and programs. The use of case studies
tends to promote application of theoretical knowledge and positive changes in affective learning, or
students’ attitudes and values.
Person-focused learning approaches move beyond case studies and directly include persons with
disabilities and family members as partners. Research and teaching-involving people with disabilities
assume that validity is strengthened through the direct involvement of people who experience
disability daily (1). Kvale and Burns discuss threats to validity and the need to reconceptualize
validity in qualitative research (2, 3). Due to the integral involvement of the researcher to conduct
qualitative research, Kvale argued that qualitative research requires attentiveness to the concept of
validity and its social construction with constant reference to the values, attitudes and experiences of
the researcher and participants (2). Further, qualitative research methodology applies to interactive
teaching, in which themes are explored and developed based on real-life scenarios (4). Participatory
action research, a qualitative research process, directly involves key stakeholders in all phases of
investigation (5, 1). In the present study, partnerships with persons with disabilities and family
members began and continued throughout the design, implementation, and evaluation of co-teaching
activities.

Corresponding author: Nancy B. Robinson, CSU Chico, SPPA Program, AJH 107B, Chico, CA 95929-0350, 530 898-6848
(voice), 530-898-6612 (fax), [email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The goal of the present study is to the person-focused learning model is highlighted
demonstrate and evaluate an interactive teaching as the focus of this study and context for
method that directly involves people with participatory action research.
disabilities and their family members and the Problem-Based Learning. As stated above,
impact of this model on students’ attitudes and PBL began within medical education to increase
values, or on affective learning. Although the the application of medical theory and information
use of case study approaches in college level with specific patient case studies and has since
teaching, particularly with persons with extended to nursing, occupational therapy, and
disabilities, produces positive student learning other fields (7-11). Cockrell, Hughes, Caplow,
outcomes, the differences in approaches to the and Donaldson described problem-based learning
uses of case studies are not explored. as a “collaborative learning approach” (12).
Specifically, the researchers sought to examine Collaborative learning is premised on Vygotskian
the effectiveness of person-focused learning to concepts that define learning as the social
promote the responsible conduct of research construction of knowledge. The cooperation and
among graduate, post-graduate and doctoral shared resources that take place in PBL learning
students. reflect tasks in “real world” settings. These
Three major developments in policy, authors outlined six basic phases in PBL:
program development and teaching practices led (a) encounter with the problem; (b) free inquiry;
to the development of person-focused learning. (c) identification of learning issues; (d) peer
First, shifts in legislation and policy began in the teaching; (e) knowledge integration and
1950’s and 1960’s in the US, which continues (f) problem resolution. Based on their
today with increasing emphasis and advocacy for investigation of student’s perspectives of PBL,
the rights of people with disabilities to have Cockrell et al. found three key areas of student
equal access to all arenas of community life. perspectives of PBL: ownership, group
Second, increasing focus on rights and advocacy dynamics, and tutor feedback (12). Students
for people with disabilities contributed to the reported a deeper level of understanding and
self-determination movement that places retention in the PBL process compared to more
decision-making and life choices with the people traditional teaching approaches and increased
affected, people with disabilities. Third, teaching awareness of team building skills. Students stated
practices in higher education shifted from a preference for tutors who were non-directive
traditional didactic models to interactive, and non-obtrusive. Students reported that the
problem-solving models that strive to establish benefits of collaborative learning included: a)
critical thinking skills among students in learning to become part of a learning community,
preprofessional training programs. The combined and b) learning to speak the language of the
influences of these broadly defined trends in community of professionals within the discipline.
policy, program, and professional practice are Inquiry-based learning. Inquiry-based
particularly relevant in higher education, where learning (IBL) uses a case-study process to
the forming of future professionals’ values, encourage student responsibility for learning
attitudes, knowledge, and skills are critical for outcomes. Inquiry-based learning is similar to
future practice and partnership with people with PBL in teaching methodology and includes
disabilities. presentation of case studies and the application of
Teaching methodology in professional a problem-solving process that students use to
training programs is changing from a didactic identify relevant issues that require further
approach to an interactive model that requires research. However, rather than resolving the case
students to take responsibility for their own through a diagnosis, IBL focuses on the inquiry
learning (6). Medical education first developed process using issues that are relevant to the case
problem-based learning (PBL) to create a student (13, 14). As in PBL, students take ownership
driven learning model. PBL was since adapted to from the beginning, as in PBL and work in small,
curricular content in several health, human tutorial groups guided by a faculty member. The
service, and education disciplines. Beginning case is discussed and analyzed based on what
with PBL, four approaches to interactive and information is known, further information
problem-solving approaches to teaching are needed, and the identification of learning issues
briefly described in this paper. The strengths and that require further research. The cases provide a
contributions of each model are addressed and structure and format that guide students to
118
––––––––––––––––––––––––––––––––––––––––––––––– Robinson & Sadao, Research Integrity and Persons with Disabilities
explore potential solutions to posed problems. and individuals with disabilities attend the
Casebooks are now an accepted technique in closing session for the purpose of providing
preservice teacher training programs (15). As is feedback to students on the scope of their work,
indicated in PBL, the use of a case encourages relevance to their particular case, and quality in
group work that inevitably models collaborative addressing the particular issue selected. As in the
communication skills found in the field. The IBL closing session, faculty assist students in
paper case leads learners to apply skills learned summarizing their analyses their individual
to field projects (16). Students then conduct research and relate students’ findings to broad
independent research and at a later session, issues affecting families and persons with
present the results of their research that disabilities.
originated from the initial case study. Faculty Person-focused learning. Person-focused
members with the focus on critical analysis of learning (PFL) incorporates teaching and
relevant policy, program, advocacy, financial, learning methods included in the previous
cultural, facilitate summary and wrap-up models, but builds on elements found in each
discussion and community issues related to the preceding approach. The elements of problem-
case. solving and critical thinking that are hallmarks of
Family-focused learning. Family-focused PBL and IBL approaches are also essential to
learning (FFL) formed in the context of person-focused approaches. As in the FFL
interdisciplinary education for health model, person-focused learning is designed and
professionals to provide a model of direct implemented with the participation of families
involvement of family members in the teaching and persons with disabilities. A new element is
process (17). Family-focused learning follows the service-learning aspect of PFL. In the PFL
the inquiry based approach through a series of approach, students are required to complete a
sessions that begin with identification of issues project that responds to needs and concerns
around a particular family with an individual identified by the family or individual (18). The
member with a disability, and close with student involvement of persons with disabilities,
presentation of research issues related to the families, faculty, and students in the development
particular family that is participating in the and implementation of the teaching experience
teaching and learning process. The key produces a qualitative shift in teaching
difference in the FFL, compared to the previous methodology and creates an action research
models described, is that actual families and model (4, 19-21). In the case-study approach,
people with disabilities participate in the teaching students respond to the issues presented for the
process with faculty, interact with faculty and primary purpose of advancing their own learning.
students throughout the development of case In the person-focused model, students are placed
information to be presented and provide in an interactive relationship with family
supportive critique to students in their work. members and individuals from the outset of the
Similar to PBL and IBL, the FFL model requires experience. The student learning goals, from the
an initial session to present concerns and faculty perspective, involve: a) application of
information that guide student inquiry. In theoretical knowledge with real families and
contrast to the other two models, FFL involves individuals with disabilities; and b) development
actual family members who present the “family of resources that respond to the needs expressed
story” to students through video and written by families and individuals.
media. The development of the video is a joint In the current study, the authors were
venture for the family and participating faculty concerned with the qualitative impacts of the
members that can require two or more sessions. PFL model on the people involved: students,
When the family is satisfied with the video families, and persons with disabilities. The
presentation, the tape is shared with students of unique features of the PFL model which
several health, human services and education incorporate problem solving in a real-life context
disciplines that identify key issues in a problem- and service to families and individuals require
solving process similar to the two models already systematic evaluation. The assumption that direct
described. Following completion of independent involvement of actual family members and
research, students prepare issue papers and people with disabilities increases validity and
present them to the family and/or individual for thus applicability of the teaching process
critique in a closing session. Family members required empirical investigation and
119
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Disability Studies
IHE CSU Chico, CA UOP, Stockton, CA
University of Hawaii
Speech Pathology: Special Education: Disability Studies: Team
Course
AAC Methods Work
Upper Division & Upper Division/ Upper Division &
Level
Graduate Graduate Graduate
Interdisciplinary
Dept. Speech Pathology Special Education
Disability Studies
Students 18 students 40 students 13 students
Table 1. Student participants in Person-Focused Learning at three universities.

consideration of the ethics involved. In this outcomes were identified through semi-
study, the authors sought to systematically structured interviews completed with the
evaluate the reciprocal impact of interactive investigator.
teaching on student learning outcomes and The courses that provided the context for the
people with disabilities, specifically with people study included a core special education course,
with disabilities in direct interaction with an elective course in augmentative and
students for the duration of semester-long alternative communication (AAC), and
courses. interdisciplinary teamwork course. Family
The foci of investigation centered on three members and individuals with disabilities
questions: participated as teaching partners with faculty
1. What are student perceptions of the PFL members. Courses were located at California
process, both in the process of interacting State University, Chico; the University of the
with families and individuals and in learning Pacific in Stockton, California; and the
University of Hawaii. Students who participated
outcomes?
in the courses included three groups, shown in
2. What are family member and individual Table 1.
perspectives of the PFL process, regarding Characteristics of the seven individuals and
their partnership role in teaching students families who participated in the study are listed
and project outcomes? below:
3. What are ethical and logistical consider- •Three adults, three children
ations for the replication of PFL in human •Communication disorders and physical
service training programs, particularly disabilities in all subjects
related to disabilities? •Two individuals with Asian/Pacific Islander
ethnicity
Methods •Five individuals were Caucasian
The study was completed in the context of three Course content and learning objectives differed
interdisciplinary courses at three different across the three sites. However, key variables
university sites, with 71 students and 7 families were held constant in teaching methodology. All
including persons with disabilities. While course courses included persons with disabilities and/or
content differed across the three sites, teaching family members who participated in the design
methods were similar. Teaching partnerships and implementation of the curriculum. The major
used principles of “Family Centered Care,” in requirement in each course included direct
which family concerns drive professional interaction with persons with disabilities and
interventions (22, 14, 23). Key steps in the family members in the design and development
teaching partnership included: (a) determination of adaptive equipment or technology to meet
of family priorities; (b) adaptations to meet needs identified by the individual and family.
family and individual needs; (c) family input in Students engaged in a common process that
project development; and (d) evaluation of included identification of needs by persons with
completed projects by family members and disabilities and/or family members adapted from
persons with disabilities. Student learning participatory action research (5, 1). Eight steps
outcomes were evaluated with qualitative surveys were completed in the person-focused learning
completed independently. Family and individual teaching process. First, faculty developed

120
––––––––––––––––––––––––––––––––––––––––––––––– Robinson & Sadao, Research Integrity and Persons with Disabilities
curriculum information about individuals in investigative and problem-based learning in
partnership with identified families and persons direct interaction with people with disabilities
with disabilities. Second, students reviewed and family members. Analysis of student
available information about the family and/or surveys identified seven themes: (a) attitudinal
individual determine an initial developmental or change; (b) authentic engagement; (c) critical
environmental concerns identified by the family thinking; (d) sensitivity to families and
and/or individual. Third, student groups individuals; (e) collaborative teamwork;
conducted brainstorming regarding potential (f) preparation for inclusion; and (g) self–
family and individual concerns. Fourth, students efficacy/skills to adapt materials. Examples of
prepared interviews based on guidelines provided student comments are included below related to
by faculty. Fifth, students conducted interviews each theme:
with individuals and/or family members. Sixth,
the working group met to identify adaptation or Attitudinal Change.
“There are many things that disabled students
support project based on results of prior
are able to do…most important to focus on
information and interviews with individual and those strengths.” 18c
family members. Seventh, student groups “I realized how many aspects of a person’s life
presented completed projects to individuals and can be affected by a disability.” 18c
family members. Finally, student evaluations of “It made me realize how difficult it must be to
the process and projects were completed. have a child with a disability, or to be a child
The qualitative effectiveness of the person- with a disability; everyday actions are so
focused learning process was evaluated by: difficult!” 19c
(a) student perceptions of learning outcomes; and “I find myself constantly looking at isles in
(b) perceptions of family members and persons stores, toys, elevators, etc. to see how they
with disabilities. Methods of evaluation included could possibly be adapted to better suit the
needs of children with disabilities—more
student’s self reports and family/individual
awareness.” 7c
interviews. “I think it helped me look at adapting
Self-Report. Students were requested to equipment as a fun responsibility instead of a
complete qualitative comments in response to required duty.” 8c
questions designed by the investigators. “It has helped me to realize that children with
Questions addressed students’ perceptions of the disabilities have a vast amount of needs, and
learning process and outcomes related to direct that each child’s needs are unique. Adapted
interaction with family members and persons equipment may still need further adaptations
with disabilities. to meet a specific child’s needs.” 10c
Family/Individual Feedback. Individuals Authentic Engagement.
with disabilities and family members were asked “The hands-on work helped me to develop a
to evaluate their participation in the courses in a better understanding of a family’s needs and
teaching/consultant role. Perceptions of these wishes for their children. Though most of
participants were also requested regarding the all…learning the true-to-life reality of the
quality of student projects and interaction with processes involved in working with a family.”
family members and persons with disabilities. 12c
As the focus of teaching included adaptations and “Actually making the adaptations brings more
assistive technology, participants were requested involvement and thus more interest, which lead
to evaluate benefits and changes related to to more learning.” 12c
“I think with the case study, it is each to
adaptations or resources developed by students.
maintain the same frame of reference and not
to expand on ideas or think about new things.
Results and Discussion With the adapted equipment, new ideas or
Results of the study are discussed in relationship problems are presented and we brainstormed.”
to perceptions of student learning outcomes and 10c
impacts on family members and persons with
disabilities. Critical Thinking.
“This assignment makes you think about
Student Problem-Solving. Student responses
aspects of disabilities that normally one
to qualitative questions were analyzed to wouldn’t consider.” 2c
determine recurring themes related to “We had discussed the written assignment a lot,

121
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
even before we knew what the questions were. individual.” 15c
We were always thinking, how it would help
B.” 6c Self-Efficacy and Adaptive skills.
“The most important part of this assignment
Sensitivity to Families and Individuals. was that it opened a door for me and pretty
“Meeting in an informal setting allows both much told me that I had the potential to help
sides of the team to get to know each other with any child with a disability.” 3c
out the pressure of a meeting…with the family “I learned that I take my skills and abilities for
relaxed we can start building relationships.” granted. From meeting B., I realized that many
16c aspects of daily living would be difficult for
“Getting to know the family was an important her, and in order for them to function at her
milestone for us.” 16c level, more things would need to be adapted.”
“It has made me realize that the parents are very 10c
important in identifying the a child’s needs.” “Yes, because it provides hands on time that I
16c will remember more than any case study. It is
“I thought it was very useful to ask T. [the also more fun than any case study.” 9c
parent] our questions because we had to know “I liked the developmental framework and the
exactly what her situation was so the outcome way this was all set up. It was very realistic to
would be helpful.” 5c what we deal with in our real jobs and it was
very hands on.” 20c
Collaborative Teamwork. “It makes me become more aware of the types
“Yes, because we need each other’s specialized of things; a lot of things that I would have never
skills along with knowledge and creativity.” thought of.” 13c
14c Family and individual interviews revealed four
“It was a great idea to work in a group because themes: (a) interaction with students; (b) self-
everyone has different ideas which we can
validation; (c) support networks; and
bring together. Then everyone has different
talents which were utilized in the production (d) alternatives to meet individual needs.
process.” 12c Families and individuals commented that they
would participate again. Table 2, below
Preparation for Inclusion. demonstrates representative feedback provided
“This is something I will have to do in my by family members and person with disabilities.
classroom so I appreciate the preparation.” 2c Ethical issues identified included the need to
“To find different ways to teach someone the
(a) respect individual choice in participation;
ABCs and how slow the song needs to be so
that the child can learn.” 9c (b) confidentiality; (c) honor individual priorities
“It has made me realize that each child with a and (d) respect family differences. Comments
disability is an individual; helping each child provided by families and individuals at the
can be done only if that child is looked at as an completion of each class indicated the possibility

Theme Identified Family Comments

“Having students come to our home was a highlight of the week for J., he
Interaction with students looked forward to it all week.”
“Students gave S. attention and made us appreciate his importance.”
“I am getting braver to ask fo r what my son needs.”
Self-validation “I always knew that J. knows more and the students helped to d ocument
that.”
“It is wonderful for our whole family to participate with the
Support networks students…going to the beach together was a first for us.”
“All of the time and support has given S. a chance to get out more.”
“The help provided by the class gave S. a way to communicate that he did
not have before.”
Alternatives to meet needs
“We want S. to learn with the other kids and he shows the book to every
one who comes over.”
Table 2. Qualitative themes and family comments regarding Person-Focused Learning Outcomes.

122
––––––––––––––––––––––––––––––––––––––––––––––– Robinson & Sadao, Research Integrity and Persons with Disabilities
of initial reluctance to participate. One parent behavior with human subjects in the career
commented that she initially was nervous when contexts of service and research.
meeting the students for the first time, The qualitative evaluation results of student
particularly due to cultural differences between learning outcomes indicate that involvement of
them. However, this parent later reported that persons with disabilities in the teaching process
her feelings changed later after realizing how provides authentic learning that cannot be
much attention and support the students replicated with more traditional didactic
demonstrated toward her son. This mother’s methods. Further, involving family members in
comment highlights the need to honor individual the teaching and evaluation process at all levels
family priorities that may be based on cultural follows a participatory action research process
styles, educational background, language and allows “checkpoints” for subjects to be fully
differences, and other variables. Related to this is cognizant of the research agenda and purposes.
the need to respect and understand family Thirdly, including people with disabilities in the
differences and follow the lead of the individual research/teaching process strengthens validity as
or family to determine the most appropriate time recommended by Kvale and Burns (2, 3).
and place to conduct interviews and project Further, reciprocity in the learning setting is
activities. achieved where students learn the needs of
The results revealed positive qualitative families and the value their knowledge when
student learning outcomes. People with designing materials and technologies to assist
disabilities and family members reported that them in the learning environment. The research
their participation provided important benefits participants are valued by the researchers and the
that included perceptions of increased self- students involved in the assignment and the
efficacy and competence when interacting with student-made products are valued by the families.
students. Risks were not specifically identified The demonstration of a pre-service training
by families or persons with disabilities, but approach that teaches reciprocal relationships
inferred from their feedback. The responsibility with subjects is perhaps the key finding with
to consider risk, which may include risks to implications for training future professionals in
privacy of participants, remains with the the area of responsible conduct of research. Not
researcher who embarks on teaching partnerships only did students demonstrate qualitative
with families and persons with disabilities. evidence of critical thinking in the learning
Comments provided by students in all thematic process, the direct interaction with subjects in the
areas reported revealed increased awareness and action research model employed in Person-
respect for the life experiences of persons with Focused Learning showed an effect on the
disabilities and family members, thus students’ sensitivity toward persons with
establishing a foundation for ethical behavior in disabilities and family members. The
future professional roles with persons with demonstrated effect on students’ sensitivity with
disabilities, including teaching, service, and subjects could effect future professional ethics
research. and conduct. While, further study is needed to
determine attitudes and values that are directly
Summary related to the responsible conduct of research
The results of the present study support the with human subjects, student attitudes toward
effectiveness of interactive teaching, specifically subjects are considered a critical variable of
Person-Focused Learning, to promote student ethical behavior. The question of what particular
learning outcomes that demonstrate respectful teaching model effectively trains professionals
and responsible professional attitudes and who are prepared to implement responsible
behavior with persons with disabilities and conduct of research was only partially addressed
family members. The specific student learning by the present study. The attitudes and skills
outcomes were found in both cognitive and required for responsible conduct of research are
affective domains, as seen in students’ clearly a constellation of knowledge and ethics
evaluations of the learning experience. These that require further explication.
findings have implications for preservice training This qualitative study explored person-
of health, human service, and education focused learning principles in several preservice
professionals to establish a foundation for ethical courses and revealed positive findings for
students and the families who shared their
123
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
stories. The “realness” of the learning setting 9. Bridges, EM. & Hallinger, P. Problem-based learning in
allowed researchers to identify multiple learning leadership development. 1995; Portland: University of
outcomes and ethical issues when involving Oregon. ERIC Clearinghouse on Educational
people with disabilities in a teaching setting and Management.
10. Gist, GL. Problem-based learning: A new tool for
research endeavor. Bowen identified the need to environmental health education. Journal of
strengthen internal validity through the Environmental Health 1992; 54: 8-13.
integration of qualitative and quantitative 11. Royeen, C.B. A problem-based learning curriculum for
research methodology (24). Further research in occupational therapy education. Journal of
PFL is needed to a) specify affective and Occupational Therapy 1995; 49: 338-346.
cognitive student learning outcomes; b) quantify 12. Cockrell, K.S.; Hughes Caplow, J.A.; Donaldson, J.F. A
changes in student attitudes; b) compare PFL context for learning: Collaborative groups in the
teaching to other problem-solving approaches; problem-based learning environment. The Review of
c) identify long range impacts on student Higher Education 2000; 23: 347-363.
13. Geln, S. Towards a new model of nursing education.
learning; d) develop guidelines for replication; Nurse Education Today 1995; 15: 90-95.
and e) explore the use of PFL to teach 14. Matteoli, RF.; Verhey, MP. & Warda, MR. Collaboration
responsible conduct of research. The between families and university faculty: A partnership in
philosophical attitude and the research model in education. Journal of Psychosocial Nursing 1994; 32:
the present study provide a framework for 17-20.
preservice education and further research to 15. Shulman, J., Lotan, RA., & Whitcomb, JA. Groupwork
determine specific professional attributes that in diverse classrooms: A casebook for educators. New
lead to affective, cognitive, and ethical York: Teachers College Press; 1998.
foundations for the responsible conduct of 16. Boyle, JR. & Danforth, S. Cases in Special Education.
Boston: McGraw-Hill. 2001.
research, particularly with persons with 17. Ratliffe, K., Stodden, N., Guinan, M., & Robinson, N.R.
disabilities. Family partnerships in teaching: family-focused
learning. Journal of intellectual disability research 2000;
Bibliography 44 (3 & 4): 416.
1. Guerrero, JL. Research paradigm shift: Participatory 18. Stevens, L. Academic Service Learning within an AAC
action research. Working Paper No. June 1995. Curriculum. 2000 Poster presented at American Speech,
Rehabilitation Research and Training Center of the Language and Hearing Association Annual Conference,
Pacific. San Diego State University, Interwork Institute. Washington, D.C., November 16, 2000.
2. Kvale, S. The social construction of validity. Qualitative 19. Henry, SK., Scott, JA., Skobel, B., Jones, A., Cross, S.,
Inquiry 1995; 1 (1): 19-40. Butler, C., & Blackstone, T. Linking university and
3. Burns, PA. Securing internal validity: How to avoid the teacher communities: A Think Tank Model of
threats [serial online] 1997 [cited 2000 Nov] Available professional development. Teacher Education and
from: URL:http://trochim.human.cornell.edu/tutorial/ Special Education 1999; 22: 251-268.
burns/int.html 20. Hutchinson, NL. & Martin, AK. Fostering inclusive
4. Newman, JM. Action research: A brief overview [serial beliefs and practices during preservice teacher education
online] 2000 Jan [cited 2000 Nov] Forum Qualitative through communities of practice. Teacher Education
Sozialforschung / Forum: Qualitative Social Research and Special Education 1999; 22: 234-250.
[On-line Journal]. Available from: URL: http:// 21. Perry, NE., Walton, C. & Calder, K. Teachers
qualitative-research.net/fqs. developing assessments of early literacy: A community
5. Whyte, WF. Participatory action research. Newbury of practice project. Teacher Education and Special
Park (CA): Sage Publications; 1991. Education 1999; 22: 218-233.
6. Casey, MB, & Hwoson, P. Educating preservice 22. Harry, B., Rueda, R., Kalyanpur, M. Cultural
students based on problem-centered approach to reciprocity in sociocultural perspective: Adapting the
teaching. Journal of Teacher Education 1993; 44: 361- normalization principle for family collaboration.
369. Exceptional Children 1999; 66:123-136.
7. Chabon, S. & Wilkerson, DL. Discussions on diversity: 23. Gartner, A., Lipsky, DK., Turnbull, A. Supporting
A problem-based instructional model to facilitate families with a child with a disability. Baltimore: Paul
cultural competence. Poster presented at American H. Brookes Publishing; 1991.
Speech, Language and Hearing Association Annual 24. Bowen, K. The Sin of Omission -Punishable by Death to
Conference, Washington, D.C., November 16, 2000. Internal Validity: An Argument for Integration of
8. Branda, LA. Implementing problem-based learning. Qualitative and Quantitative Research Methods to
Journal of Dental Education 1990; 54: 548-549. Strengthen Internal Validity[serial online] 1996 [cited
2000 Nov]. Available from: URL: http://
trochim.human.cornell.edu/gallery/bowen/hss691.htm.

124
4. Conflict of Interest
What is Driving Policies on Faculty Conflict of Interest? Considerations for
Policy Development
Mildred K. Cho, Center for Biomedical Ethics, Stanford University, USA
Ryo Shohara, Institute for Health Policy Studies, University of California, San Francisco,
USA
Drummond Rennie, Institute for Health Policy Studies, University of California, San
Francisco, USA

Keywords: Academic policies, Academic-industry ties, Conflict of interest, Research integrity

There are several factors driving policies on conflict of interest of faculty at academic research
institutions in the United States today. The first is that researchers and institutions have a greater
number, and a wider variety of financial conflicts of interest, especially in the area of biomedical
research. Sometimes, these financial interests appear to lead to very bad outcomes, and when that
happens, public scrutiny of the financial interests increases. Sometimes, this leads to new policy.
What is the current state of academic-industry ties in biomedical research? In 2000, the NIH’s
budget is $17.8 billion (1), while the pharmaceutical industry’s R&D budget is $22.4 million (2).
Krimsky found that 34% of research articles published in the top 14 biomedical research journals in
1992 had undisclosed financial ties of a lead author. These ties included holding a patent on an
invention related to the published research, or being on an advisory board or a major shareholder in a
company whose activites were related to the published research (3). In a review of FDA records,
USA Today reported that 54% of the time, experts hired by the FDA to advise on safety and
effectiveness of drugs have a direct financial interest in the drug or topic they are asked to evaluate
(4). Therefore, academic-industry ties are now the norm, rather than the exception.
Academic-industry ties have been the apparent cause of bad outcomes, including censorship of
data (5, 6), publication bias (7-10), lower quality of research (11), and harm to research subjects,
including death (12). Although it is impossible to determine a causal link between financial interest
and adverse outcome in individual situations, systematically gathered evidence suggests that, in the
aggregate, academic-industry ties can have adverse effects on the scientific process and outcome in
the aggregate (13).
One bad outcome in particular has led recently to public scrutiny and re-examination of policies
on conflicts of interest — the death of Jesse Gelsinger, who was a research subject in a Phase I
clinical trial of gene transfer at the University of Pennsylvania (12). Much attention focused on the
financial ties of investigators and the investigators’ institution with a company that was, in part,
sponsoring the trial. Although, again, it is impossible to prove that there was a causal link between
the financial ties and the death of Mr. Gelsinger, it was a link that was inevitably made, time and

Corresponding author: Mildred K. Cho, PhD, Stanford Center for Biomedical Ethics, 701A Welch Road, Suite 1105, Palo
Alto, CA 94304, 650-725-7993 (voice), 650-725-6131 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
again. A quote from a recent newspaper article institutions whose faculty apply for PHS or NSF
sums up the public perception: funding develop and implement their own written
Paul Gelsinger, Jesse’s father, said yesterday rules for faculty conflicts of interest. These
he had undergone a painful change of heart in institutional policies must conform to, but need
the year after his son’s death, at first fully not be limited to, federal regulations. Indeed, the
trusting the researchers and holding them majority of institutional policies go beyond
blameless and then gradually, as disclosures of federal regulations in scope and management of
apparent wrongdoing emerged, concluding that
he had been duped by scientists who cared more
conflicts of interest, but most do not state
about profits than safety. (14) specific limits on financial interests, even when
After Mr. Gelsinger’s death, the National in conjunction with company-sponsored research
Institutes of Health (NIH) held a public meeting (16). Most of these policies imply or state that
this year to re-examine some aspects of conflict conflicts of interest are dealt with on a case-by-
of interest policy, and several professional case basis, and seem to rely heavily on disclosure
organizations, including the National Academy as a primary mechanism for dealing with conflict
of Sciences, the American Association of of interest.
Medical Colleges (AAMC), and Association of Some research journals have developed
Academic Health Centers (AHC), the American policies that require disclosure of authors’
Association of Universities (AAU), and the financial interests to editors and reviewers .
American Association of University Professors However, such disclosures often do not surface
have all assembled internal groups to do the on the pages of the published articles, so their
same. effects are limited (Krimsky, this volume).
The AAMC, AHC, and the AAU created
What are the current policies on faculty guidelines for faculty conflict of interest long ago
(17-19), and although they thoughtfully outline
conflict of interest?
policy considerations, they are not specific and
Current policies on faculty conflict of interest
are not enforced. Finally, in the wake of Jesse
exist at several levels, including federal, state,
Gelsinger’s death, two professional societies (the
institutional regulations, editorial policies at
American Society of Gene Therapy and the
research journals, and statements by professional
American Society of Human Genetics) have put
societies. All are limited, however, in different
forward statements that faculty having financial
ways. The most widespread federal rules include
interests in companies sponsoring their gene
the “Objectivity in Research” regulations (15).
transfer research is inappropriate and should be
These are applicable only to researchers who
avoided (20, 21). These statements only apply to
apply for research funding from the National
gene transfer research, however, and also have no
Science Foundation and the Public Health
enforcement power.
Service (PHS), which includes the NIH. These
regulations are limited to disclosure of financial
ties that could be construed to affect the publicly- What should we do about conflicts of
funded research, and to financial ties that exceed interest?
$10,000 annually or 5% equity interest. Thus, The answer to the question, “what do we do
financial ties in the context of industry-funded about conflicts of interest?” depends upon the
research, where more serious conflicts of interest answers to the questions, “what is conflict of
might be found, are not covered under these interest?”, “what is the primary interest of
regulations. academic institutions and the government?”, and
In addition to federal regulations, there are “what are the secondary interests we are
state laws that might apply to faculty at public concerned about?”
institutions. For example, some states prohibit or What is conflict of interest? Opinions are
require full disclosure of gifts to public diverse. Many make the distinction between
employees, which include faculty of state “actual” and “potential” conflicts of interest.
universities. These state laws often do not apply Others call it scientific misconduct (22).
to private universities, and are not uniform from Depending on how one defines conflict of
state to state. interest, one may be led to base policy on
Institutional policies are mandated by the evidence of bad outcomes or on ethical or
federal regulations, which require that professional values. We define conflict of

128
–––––––––––––––––––––––––––––––––––––––––––––––– Cho, et al., What is Driving Policies on Faculty Conflict of Interest?
interest as the co-existence of a primary interest the characteristics of a situation rather than the
or duty (such as research integrity, patient outcome.
welfare, or education) and a secondary interest What is the primary interest? Lack of clarity
(such as financial gain or recognition) (23). The about the primary interests of researchers and
policy concern is that the secondary interest their institutions will lead to bad policy, because
exerts undue influence on the judgements made one of the points of having the policies is to
in the course of executing the primary interest, protect the primary interests. So, the question is,
leading to adverse outcomes (such as research what are the roles of academic institutions and
bias or adverse effects on research subjects). the government in the conduct of science? The
It is important to remember that conflict of passage of the Bayh-Dole Act gave the
interest rules are activated in the absence of a government a new role in academic research,
“crime” (24). Stark likens them to speed limit namely, “to promote the marketing of inventions
laws. In contrast to laws against murder, which developed under federally supported research and
are aimed at activities that, in themselves, are development projects by nonprofit organizations
deemed immoral and are not in the public and small business firms.” (27)
interest, speed limit laws are aimed against Government specifically encouraged
conditions that predispose to the activities that academic institutions to be involved in the
are not in the public interest. So, while driving at marketing of inventions. Universities have taken
70 miles per hour may not in itself be wrong in this encouragement to heart, “… shifting from
the way that murder is wrong, high-speed driving ivory tower to revving economic engine.” (28)
may enhance the chances of causing harm to The new role of universities as economic engines
others. Some drivers might be quite capable of leads to expectations that they create jobs and
avoiding crashes at even 200 miles per hour, but even whole industries. In fact, the government
because it would be difficult and impractical to has implicitly adopted the values of the business
determine who they are and whether they are so world, where money is an incentive for
capable under all circumstances, the laws are employees to work in the interests of
aimed at preventing the situation rather than shareholders. In this model, the secondary
particular outcomes. However, there may be (financial) interest is considered to be in
certain speeds that would be considered alignment with the primary interest, rather than
“reckless” in almost any circumstances, and thus acting as a competing interest. By contrast, the
immoral – and there may be analogous financial model of professionalism says that the Bayh-
interests. Dole Act and related legislation specifically put
However, there is an important difference not only faculty but institutions in a position of
between speed limit laws and conflict of interest conflict of interest. If academic institutions and
regulations, in that speed limit laws apply to their faculty are expected to add economic goals
everyone, whereas conflict of interest laws apply to their primary missions, can those institutions
to groups that have a fiduciary relationship to the be expected to be effective at developing and
public, such as public officials or professionals. enforcing conflict of interest rules for their
This distinction is important, because it means faculty? This seems to be a dangerous thing to
that there are reasons to set the rules by criteria ask.
other than probability of harm to the public, We must be clear about whether academic
namely in order to earn or preserve the right to institutions should take on economic health as a
occupy the special position in society (25). primary interest. We must also be clear about
This definition of conflict of interest implies whether we are concerned only with or more
that there can be no distinction made between concerned about certain kinds of primary
“actual” and “potential”. The conflicting interests. For example, is only federally-funded
interests simply either exist or they do not. They research of concern, or all research? That is,
are, in themselves, not scientific misconduct, should policies be directed only at interests that
although they may lead to misconduct. The conflict with government-funded research, or
current definition of scientific misconduct carries should they also be directed at interests that
with it the notion of wrongdoing with intent (26), conflict with industry-funded activities, too?
which is based on the proven existence of a bad Finally, we should also ask whether clinical
outcome, and is therefore incompatible with a research is of more concern than other research.
definition of conflict of interest that is based on There are good ethical reasons to distinguish
129
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
research that involves human subjects from other five methods seem to be organized smoothly
research, primarily that human subjects are along a continuum of stringency. However,
subjected directly to risks from the research closer examination reveals that there is actually a
itself. qualitative difference between these strategies,
What is the secondary interest? Lack of because they are based on different assumptions.
clarity about the secondary interests that are of In theory, all of these methods act by
concern will also lead to bad policy. Current modifying the conflict of interest situation
regulations focus on financial interests, rather through either the primary or secondary interest.
than other, less-tangible interests such as However, disclosure is distinct from all the other
academic recognition and fame, or personal ties. methods. It is supposed to act not by virtue of
This is appropriate for the time being, not supplying information to the disclosee, but
because the intangibles are less damaging, but because the release of this information is
because the financial interests are avoidable and supposed to make the discloser more aware of
because avoiding them is consistent with the role the potential effects and thus affect the
of a professional, and enhances public trust. discloser’s behavior (24). Clearly this is a weak
Financial interests have also increased to a high method because of its indirectness. In practice,
level and are deserving of attention merely the information rarely gets out to a wide
because of their frequency. Furthermore, those audience, and the discloser knows it, limiting
who point to the unfairness of concern about effectiveness. More importantly, this method
financial interests seem to imply that financial allows the discloser to feel that the act of
interests merely replace the non-financial disclosing has let him or her off the hook, and
interests, so that there is no need for special places the burden of management on the
consideration of the financial interests. However, disclosee. Stark points out that disclosure is
the literature suggests that the effect of financial based on a model where the role of the discloser
interests on biomedical research can be detected is as an “agent”, or delegate, rather than a trustee.
as an independent factor, above the background By this model, the disclosee is assumed to have a
“noise” of the want for academic recognition and large degree of control over the activities of the
fame (assuming that it exists uniformly among discloser.
researchers). In contrast, the other management methods
There is less clarity about what specific kinds are based on a trustee or fiduciary model. By this
of financial ties are of concern. Current model, the disclosee is assumed to have little
regulations focus on personal financial ties such control over the activities of the discloser and
as consulting fees, honoraria, royalties and equity therefore depends on the discloser to act in the
holdings. They generally do not consider best interests of the disclosee. Mediation and
company-sponsored research per se to be a abstention carry with them the notion that the
conflict of interest, but a growing body of fiduciary position is a role that can be filled by
literature suggests that industry sponsorship in interchangeable individuals. That is, the
itself biases research and publication (7-9, 13, protagonist can be replaced by a third party such
29). as an oversight committee or another researcher.
Divestiture and prohibition imply that the
How do we manage conflicts of interest? protagonist is not replaceable, and therefore the
Standard methods of managing, or mitigating, mitigation of the conflict of interests requires
conflicts of interest include (1) disclosure removal of the secondary interest.
(e.g., publication of a secondary interest), How we deal with conflicts of interest
(2) mediation (e.g., a blind trust, which puts a depends on how we view the players. Are
secondary interest under the control of a third researchers delegates or trustees? People who
party, or oversight, which puts a primary interest hold elected public office may better fit the
under the review or control of a third party), delegate or agency model, since the public has
(3) abstention (e.g., recusal from a primary the power to remove them from office if their
interest), (4) divesti-ture (e.g., removal of a performance is unsatisfactory. Researchers,
secondary interest), and (5) prohibition (e.g., however, are more like trustees (especially
permanent withdrawal from a whole category of clinical researchers) because it is understood that
secondary interests) (23). At first glance, these the public supports their training and activities to
perform tasks that others are not qualified to
130
–––––––––––––––––––––––––––––––––––––––––––––––– Cho, et al., What is Driving Policies on Faculty Conflict of Interest?
perform, and the public is not in a strong position Science and Engineering Ethics, 1996. 2: p. 395-410.
of control over these activities. The professional 4. Cauchon, D., FDA advisers tied to industry., in USA
role of scientists and clinicians is fiduciary in Today. 2000. p. 1A.
nature, and requires that public interests be 5. Kahn, J., et al., Evaluation of HIV-1 Immunogen, an
immunologic modifier, administered to patients infected
placed ahead of self-interest. with HIV having 300-549 x 106/L CD4 cell counts: a
How we deal with conflicts of interest also randomized controlled trial. JAMA, 2000. 284: p. 2193-
depends on how broadly we define the interests 2202.
and the conflicts. The goal of academic-industry 6. Rennie, D., Thyroid storm. JAMA, 1997. 277: p. 1238-
ties is to maintain the ability to conduct good 1243.
science and to enhance technology transfer for 7. Cho, M. and L. Bero, The quality of drug studies
public good, while preserving research integrity published in symposium proceedings. Annals of Internal
(including the direction of research) and, in the Medicine, 1996. 124: p. 485-489.
case of clinical research, protecting human 8. Davidson, R., Source of funding and outcome of clinical
trials. J Gen Intern Med, 1986. 1: p. 155-158.
subjects from harm. In order to achieve any of 9. Rochon, P., et al., A study of manufacturer-supported
these goals, it is essential to maintain the public trials of nonsteroidal anti-inflammatory drugs in the
trust and a sense of professionalism, in the treatment of arthritis. Arch Intern Med, 1994. 154: p.
original sense of the word (25, 30, 31), which 157-163.
includes strong self-regulation (32). 10. Stelfox, H., et al., Conflict of interest in the debate over
calcium-channel antagonists. New Engl J Med, 1998.
Recommendations for policy development 338: p. 101-6.
What are the implications of these definitions of 11. Rochon, P., et al., Evaluating the quality of articles
published in journal supplements compared with the
interests and conflicts of interest for policy quality of those published in the parent journal. JAMA,
development? First, conflicts of interest should 1994. 272: p. 108-113.
be defined by characteristics of situations, rather 12. Weiss, R. and D. Nelson, Teen dies undergoing
than by outcomes. This allows taking into experimental gene therapy., in Washington Post. 1999:
account professional values as well as evidence Washington, DC. p. A1.
that certain situations tend to lead to bad 13. Blumenthal, D., et al., Withholding research results in
outcomes. Second, we should not rely on academic life science. JAMA, 1997. 277: p. 1224-1228.
disclosure as a primary mechanism for mitigating 14. Weiss, R. and D. Nelson, Penn settles gene therapy suit.,
conflicts of interest. Instead, we should in Washington Post. 2000: Washington, DC. p. A4.
15. U.S. Department of Health and Human Services,
acknowledge that researchers have professional Objectivity in Research, 45 CFR 94, 1995, U.S.
responsibilities that are fiduciary in nature. As Government: Federal Register. p. 35810.
trustees, they should be trustworthy. Third, 16. Cho, M., et al., Policies on faculty conflicts of interest at
institutions should remember that institutional US Universities. JAMA, 2000. 284: p. 2203-2208.
interests play a role in individual conflicts of 17. AAMC Ad Hoc Committee on Misconduct and Conflict
interest, as well as the administration of policies of Interest in Research, Guidelines for faculty
about individual conflicts of interest. Therefore, involvement in commercially supported continuing
institutions should not use policies only as medical education. Academic Medicine, 1992. 67: p.
administrative tools, but also as mechanisms for 617-621.
18. AHC Task Force on Science Policy, Conflicts of interest
communicating institutional values to the public in academic health centers policy paper #1, 1990,
(24, 31), because the nature of professionalism is Association of Academic Health Centers: Washington,
to profess a vow to place the interests of the DC.
public above self-interest (33). The goal is to 19. AAU Clearinghouse on University-Industry Relations,
provide reassurance to the public that the University policies on conflict of interest and delay of
institutions have also accepted their fiduciary publication, 1984, American Association of
role. Univerisities: Washington, DC.
20. The American Society of Gene Therapy, Policy on
Financial Conflict of Interest in Clinical Research. 2000.
Bibliography
21. The American Society of Human Genetics, Statement on
1. National Institutes of Health, About NIH: Overview, .
Gene Therapy. 2000.
2000.
22. Bodenheimer, T., Conflict of interest in clinical drug
2. Pharmaceutical Research and Manufacturers of
trials: A risk factor for scientific misconduct. 2000:
America, Pharmaceutical Industry Profile 2000, . 2000.
Bethesda, MD.
3. Krimsky, S., et al., Financial interests of authors in
23. Thompson, D., Understanding financial conflicts of
scientific journals: a pilot study of 14 publications.

131
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
interest. New Engl J Med, 1993. 329: p. 573-6.
24. Stark, A., Conflict of interest in American public life.
2000, Cambridge, MA: Harvard University Press. 271.
25. Buchanan, A., Is there a medical profession in the
house?, in Conflicts of interest in clinical practice and
research, R. Spece, D. Shimm, and A. Buchanan,
Editors. 1996, Oxford University Press: New York. p.
105-136.
26. Office of Science and Technology Policy, Fact Sheet:
Research misconduct - A new definition and new
procedures for federal research agencies. October 14,
1999.
27. US Senate, University and Small Business Patent
Procedures Act, in Bill S. 414. 1979.
28. Goldberg, C., Across the US, universities are fueling
high-tech booms., in New York Times. 1999: New York.
p. A1.
29. Blumenthal, D., et al., Participation of life-science
faculty in research relationships with industry. New Engl
J Med, 1996. 335: p. 1734-39.
30. DeAngelis, C., Conflict of interest and the public trust.
JAMA, 2000. 284: p. 2237-2238.
31. Korn, D., Conflicts of interest in biomedical research.
JAMA, 2000. 284: p. 2234-2237.
32. Rothman, D., Medical Professionalism — Focusing on
the Real Issues. New Engl J Med, 2000. 342: p. 1281-6.
33. Wynia, M., Medical professionalism in society. New
Engl J Med, 1999. 342: p. 1612-1616.

132
The Commercialization of Academic Science: Conflict of Interest Policies
and the Faculty Consultant
Lisa M. Jones, Postsecondary Education Policy Studies Center, University of Minnesota,
USA

Karen Seashore Louis, Educational Policy and Administration, University of Minnesota,

USA

Keywords: Academic scientists, Conflict of interest, Faculty consultants

Social scientists have studied the effects of faculty consulting on academic productivity - teaching,
research, and service (1- 6) – and used productivity as a proxy for conflict of interest. Most recently,
writers in both the disciplinary and popular literature have addressed conflict of interest and faculty
consultants. However, little empirical research that investigates the connection between
entrepreneurial behavior, consulting, and conflict of interest, exists. This study identifies four specific
behaviors that could compromise scientific objectivity and thus, be classified as conflicts of interest:
research agenda bias, prior review, withholding, and secrecy.
These conflict of interest behaviors are grounded in the norms and counternorms of science
proposed by Merton and Mitroff (7-8). Four norms dominate the roles of scientific researchers:
universalism, dissemination, disinterestedness, and organized skepticism.
Universalism suggests that science is open to all individuals regardless of their personal traits.
The scientific method is used to pursue truth. Dissemination allows for research to become open to
all challenges, subject to verification, and widely disseminated, the antithesis of prior review.
Research advances knowledge and resides in the public domain. Results become communicated so
that others may build upon previous work to move knowledge forward. The purpose of
communication also allows for research to become open to all challenges, subject to verification, and
widely disseminated (9).
The disinterested search for truth enables scientists to explore all information regardless of where
it might lead. Science’s reliance on verification and reliability reflect institutionalized controls to
ensure that knowledge benefits humanity and allows the researchers to proceed objectively. Although
knowledge advancement is the institutionalized role of scientists, some desire credit for their
discoveries vis-à-vis election to the National Academy of Sciences or a trip to Stockholm (e.g., Nobel
Prize). Conflicts then arise over the priority of discovery that further fuels secrecy. Furthermore,
academic science is a competitive industry that encourages researchers to withhold results for
personal aggrandizement either through enhanced reputation or financial gain. Entrepreneurial
behavior is a perceived threat to the researchers’ disinterestedness in the pursuit of knowledge for its
own sake. Burton Clark views entrepreneurialism as “a characteristic of social systems...taking risks

Corresponding author:Lisa M. Jones, Ph.D., Postdoctoral Associate, Postsecondary Education Policy Studies Center,
University of Minnesota, 330 Wulling Hall, 86 Pillsbury Drive S.E., Minneapolis, MN 55455-0208, 612-625-0773 (voice),
612-624-3377 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
when initiating new business practices where the medical clinical department, one non-medical
outcome is in doubt...(10)” The scientist clinical department, and two non-clinical
maintains a vested interest in the research departments were randomly selected from each
outcomes. When individual scientists establish institution. Both the Peterson’s Guide and
research agendas based on profitability, science is University Bulletins identified 4,000 faculty that
not served. The payoff between basic research included non-clinical, clinical, and researchers
discoveries and economic profitability often funded by the Human Genome Project (HGP).
requires time that neither society nor the A stratified random sample of faculty, half of
marketplace are willing to grant academics. This whom were clinical and half of whom were non-
creates the appearance that basic research clinical, were selected from a list of faculty
projects compete with commercially viable across the 200 departments. Special provisions
proposals for funds. were made to include the HGP researchers
Finally, Merton described organized because of the broader study’s interest in
skepticism as the “temporary suspension of behaviors of genetics researchers. Ineligible
judgment and the detached scrutiny of beliefs” faculty (those who were deceased, retired, or not
that affords scientists with the opportunity to located) were omitted from the sample, leaving a
examine results using empirical or logical criteria final sample size of 3,169 faculty.
(11).
The search for truth rests upon the Data Collection
foundations of basic research. When academic The data collection process occurred from
scientists pursue lines of inquiry regardless of October 1994 through April 1995 by the Center
their commercial viability, the public interest is for Survey Research at the University of
served. However, shifting political forces place Massachusetts. Each participant was mailed a
equal or even greater importance on survey packet, which included a cover letter,
commercially viable academic science that could coded postcard, and questionnaire. The
stimulate economic growth expeditiously (12). questionnaire and postcard were to be returned
This study examines life sciences faculty separately to protect respondent anonymity.
who report earning additional income by Reminder/thank you postcards were mailed
consulting for non-profit organizations, industry, shortly after the initial mailing. Follow-up calls
and government and their engagement in actual conducted from late November to mid-February
conflict of interest behaviors. This study limits to non-respondents generated an additional 190
the definition to consulting activities for financial cases for analysis. We received useable
remuneration, and examines individuals who responses from 2,052 faculty, for a total response
select consulting as a major source of rate of 65 percent.
supplemental income from nonprofit For this substudy, the sample consists of
organizations or government agencies, private the 1,032 non-clinical faculty respondents.
enterprise, or both public and private. Selection of the individuals was assured by
Furthermore, the study examines behaviors of including only faculty who do not conduct
those who consult exclusively with one company. clinical trials on “ drugs, devices, or diagnostic or
therapeutic technologies.” The non-clinical
Methods faculty was chosen because previous research
The data source used for this study is part of the conducted using the complete sample shows that
Academic-Industry Research Relationships these individuals are on the “front end”
Study in Genetics and Other Life Sciences. The (entrepreneurial) of the commercialization
analyses here are based on data from the broader process. Furthermore, the industry relationships
study’s 1994-1995 national survey of 3,169 U.S. between clinical faculty and corporations are
faculty in the life sciences. Fifty research- structured around clinical trials rather than new
intensive institutions were selected based on the discoveries (12).
levels of National Institutes of Health funding for
1993. All medical-school departments and other Variables
academic life-science departments and graduate Faculty gender, academic rank, average annual
programs were identified using the 1994 research budget, average level of entrepreneurial
Peterson’s Guide to Graduate Programs in behavior, and average income earned above
Biological and Agricultural Sciences. One salary were used as independent variables in the
134
–––––––––––––––––––––––––––––––––––––––––––––––––––––– Jones & Louis, The Commercialization of Academic Science
statistical analyses. The entrepreneurial behavior not provide?” Yes was coded as “1” and no as
scale constructed consists of the following survey “0”.
items: “Has the research that you do at your
university resulted in....(Check one for each Statistical analysis
item)...patents applied for, a startup company.” Unless otherwise noted, statistical significance
Individuals could check either yes (coded as “1”) and the direction of reported relationships
or no (coded “0”). The next question used for between consulting and conflict of interest
this scale was: “For the firm with which you behaviors were tested by multivariate linear and
currently have the greatest involvement, which of logistic regressions. The equations were adjusted
the roles listed below do you have? (Check all for academic rank, gender, institutional control
that apply)...equity holder, company owns or (public or private), academic program ranking,
licenses a patent based on your research.” If the institutional location (metropolitan versus non-
respondent left the item blank, it was coded as metropolitan), supplemental income amount, and
“0” for no. A check mark was coded as “1” for levels of entrepreneurial behavior.
yes. The reliability for the entrepreneurial
behavior scale offered a standardized alpha of .69 Results
(n = 1032). Sixty percent (n =616) of this sample (n =1032)
report that they have consulted with either public
Conflict of Interest measures (35.2%) or private (24.5%) enterprises at least
Research agenda bias. One conflict of interest once. This contrasts with the 26% of the
measure concerns external influence on research respondents who consult with either group as a
topics: “To what extent has your selection of major source of supplemental income. Table 1
research topics been affected by...(Check one for shows the consultants’ characteristics broken
each item) a) the likelihood of commercial down by gender, academic rank, average research
application of the results.” Participants were budget, average level of entrepreneurial behavior,
offered the following response options: Not at all and average income earned above salaries.
(coded as “0”); very little (coded as “1”); to some Males account for 82% of the sample, thus it is
extent (coded as “2”); or, to a great extent (coded not surprising to see them represented more than
as “3”). The results were collapsed into a females in the consulting categories (x2 = 24.74 p
dichotomous variable coded “1” for yes and “0” < .001). Full professors represent 54% of the
for no. total sample and are also consult more than
Prior review. Another conflict of interest assistant and associate professors (x2 = 16.88 p <
measure considers the publication relationship .05). However, the assistant professors that
between faculty and the sponsor. The following consult work more with private enterprise than
item measured prior review: “Have you the public sector. One possible explanation for
personally conducted any research at your this finding is that assistant professors may have
university, the results of which are the property established relationships with companies during
of the sponsor and cannot be published without their graduate training. The results further
the sponsor’s review or consent?” Yes was coded indicate that those who consult exclusively with
as “1” and no as “0”. one company tend to be male, full professors.
Secrecy. This variable identifies the Furthermore, private enterprise consulting faculty
relationship between commercial science and have larger research budgets than non-
publication of results. “Has your university consultants, which supports a Louis et al. (13)
research resulted in findings that were never earlier study that suggested that research budget
published for proprietary reasons?” was the item reflects entrepreneurial behavior as it indicates a
used to measure secrecy. Yes was coded as “1” commitment to large-scale research. Private
and no as “0”. enterprise consultants also report more
Withholding. The final conflict of interest entrepreneurial behaviors. The analysis indicates
measure for this study considers the sharing the specific entrepreneurial activities of these
relationships between academic researchers. individuals: 65% have applied for patents (x2 =
This item asks individuals to report their denial 63.99 p < .01); 20% have started new companies
of others’ requests for research tools: “In the last (x2 = 15.19 p < .01); 23% hold equity in a
3 years, have any other university scientists company (x2 = 82.87 p < .001); and 15% are
requested any results or materials that you did involved with companies that own patents from
135
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
their university research (x2 = 31.94 p < .001). (x2 = 70.09 p < .001).
When faculty who consult exclusively with Conflict of interest variables. When
one company were compared with those who do consultants were asked to report on the conflict
not (including non-consultants), exclusive of interest variables used in this study, we found
consultants report higher levels of entrepreneurial that of those who answered “yes”, the majority
behavior, research budget, and amount earned were private enterprise consultants. Table 3
above their institutional salaries. Table 2 shows shows these results. Private enterprise and
the mean differences between these groups. nonprofit/government consultants were most
Exclusive consulting offers greater financial represented in research agenda bias (x2 = 26.58
rewards for the academic scientist, which should p < .001); prior review (x2 = 37.15 p < .001);
increase the potential for them to defy research withholding (x2 = 11.49 p < .01); and trade
behavioral norms for self-aggrandizement. secrets that resulted from university research (x2
The analysis indicates the specific = 10.61 p < .05). The results for secrecy were
entrepreneurial activities of those who consult not statistically significant.
exclusively with one company: 72% have Logistic regression analyses. Entrepreneurial
applied for patents (x2 = 30.41 p < .001); 35% behavior level (0 to 4) is associated with private
have started new companies (x2 = 33.65 p < enterprise consulting when gender, academic
.001); 35% hold equity in a company (x2 = 83.61 rank, teaching, publication numbers, service,
p < .001); and 30% are involved with companies research budget, and amount of supplemental
that own patents from their university research income are held constant. The most meaningful
Characteristics
Gender Rank Research Entrepre- Income
Budget neurial o ver
Consulting: Male Female Assist. Assoc. Full
Behavior Salary
No 79% 21% 1 3% 24% 63% 239,752 .43 4,9951
Consulting
Public 80% 20% 8% 29% 63% 355,494 .472 3,8803
Consulting
Private 96% 4% 1 7% 22% 61% 3 97,3374 1.145 1,52016
Consulting
Table 1. Consultant characteristics (N=1032) reported in percentages and means. 1 Difference between non- and public
consultants (p < .001) 2 Difference between public and private consultants (p < .001) 3 Difference between public and
private consultants (p < .05) 4 Difference between public and private consultants (p < .001) 5 Difference between non-
and private consultants (p < .001) 6 Difference between public and private consultants (p < .001)

Research Budget Entrepreneurial Behavior Income over Salary

Consulting:
Exclusive 365,5681 1.762 22,1703
All Others in Sample 269,196 .48 5,595
Table 2. Mean differences between exclusive consultants and all others in the sample on research budget, entrepreneurial
behavior, and amount earned over income. 1 (p < .05) 2 (p < .001) 3 (p < .001)

Behaviors
Research Prior Trade
Withholding** Secrecy
Bias*** Review*** Secrets*
Consulting:
No Consulting 23% 1 1% 9% ns 6%
Public 24% 9% 8% ns 7%
Consulting
Private 43% 2 9% 18% ns 12%
Consulting
Table 3. Consultant reports (N=1032) of conflict of interest behaviors. ***p < .001 ** p < .01 * p < .05

136
–––––––––––––––––––––––––––––––––––––––––––––––––––––– Jones & Louis, The Commercialization of Academic Science
variable in the equation is the private enterprise as levels of entrepreneurial behavior increase, the
consultant status (t = 9.32, p < .001), followed by odds that they define research topics according to
publication numbers (t = 4.48, p < .001). The non-research-related dynamics increase by a
strength indicates that private enterprise factor of 1.65.
consultants appear more likely to engage in The second regression tests the relationship
entrepreneurial activities than either public between consulting and prior review. The results
consultants or non-consultants. The full model, indicate that private enterprise consulting has a
which explains 15% of the variance, suggests negative effect on prior review, while
that faculty who consult with private industry supplemental income amount and level of
and who have higher publication numbers are entrepreneurial behavior has a positive effect (x2
more likely to engage in entrepreneurial = 68.16, p < .001). The probability that private
behaviors than others. enterprise consultants will publish results only
There is a modest correlation between after sponsor’s review decreases by a factor of
supplemental income and private enterprise .50. However, the likelihood of prior review
consulting (r = .32, p < .001), and exclusive increases by a factor of 1.59 for rising
consulting (r = .32, p < .001). Supplemental entrepreneurial behavior levels and 1.24 for
income amount was not regressed on consulting, supplemental income amount. Essentially, a
however, because of these correlations. The private enterprise consultant is less likely to
model, which accounts for 15% of the variance, conduct research not published without the
indicates that publication numbers, service levels, sponsor’s consent. But, increased entrepreneurial
and total research budget from all sources is behavior and supplemental income do affect
closely aligned with supplemental income prior review.
amount. The most salient independent variable is Private enterprise consulting does not appear
service (t = 5.86, p < .001), followed by to affect withholding research tools from other
publications (t = 3.73, p < .001) and overall scientists who request them in either tested
research budget (t = 3.61, p < .001). model. Faculty in private institutions are less
Correlations show weak relationships likely to withhold (by a factor of .59), while
between private industry consulting and research supplemental income increases the likelihood of
agenda bias (r = .16, p < .001), withholding (r = withholding (by a factor of 1.26). When
.09, p < .01), and prior review (r = .18, p < .001). entrepreneurial behavior level is added, the
Additionally, those who consult exclusively with negative effect of institutional control remains
one company are correlated with research agenda constant, while the supplemental income effect is
bias (r = .08, p < .001) and prior review (r = .15, slightly lessened (x2 = 34.90, p < .001). Levels
p < .001). of entrepreneurial behavior increase the chance
Logistic regressions were conducted to test that one will withhold from others by a factor of
whether or not consulting with private enterprise 1.37. The results indicate that faculty in private
affects research agenda bias, prior review, institutions are less likely to withhold from other
secrecy, and withholding. The models to test scientists even when controlling for levels of
private enterprise consulting effects included the supplemental income and entrepreneurial
following control variables: faculty attributes, behavior.
institutional characteristics, academic Finally, academic program ranking decreases
productivity measures, and entrepreneurial the likelihood that a scientists’ university
behavior levels. research results in trade secrets by a factor of .56
The first regression shows that the level of while level of entrepreneurial activity increases it
entrepreneurial behavior (x2 = 74.05, p < .001) of by a factor of 2.67 (x2 = 58.30, p < .001). This
the faculty member as well as academic program model accounts for 21% of the variability for this
ranking and metropolitan location affects variable.
whether or not they allow commercial potential The models generated to explain why some
or funding opportunities to determine their scientists conduct research that is never
research agenda. This finding suggests that published for proprietary reasons were not
faculty in highly ranked programs in statistically significant. Thus, issues related to
metropolitan areas are less likely to allow secrecy as defined in this study were not
external factors such as commercial viability and examined in this analysis.
funding to affect their research topics. However, Analyses on the effects of exclusive
137
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
consulting on the conflict of interest variables academic scientists face are complex and do not
showed results that are similar to the private allow for a simple explanation.
enterprise consultant for research agenda bias (no Despite the lack of a positive relationship
effect), prior review (negative association), and between private enterprise consulting and the
withholding (no effect). These important conflict of interest variables tested in this study,
findings suggest that even the faculty member the need to protect universities, disciplines, and
who consults exclusively with one company is the public from academic renegades remains.
unlikely to violate the research norms of the Current methods such as disclosure to both
academic enterprise. academic journals and universities provide an
important mechanism to alleviate conflict of
Discussion interest. However, these policies should be
The results do not indicate that conflicts of grounded in conflict of interest behaviors, rather
interest occur with any significant frequency; to than potentials, and enforced by individuals in
the contrary, the results show that academic the academic community. Emanuel and Stein
scientists are able to balance their reported that one out of three authors of journal
institutionalized scholarly roles with commercial articles held financial stakes in reported research
science. Faculty remain embedded in their own outcomes and failed to disclose such in their
social organizations which in the case of the publications (17). If self-regulation of the
consultant includes the university, the discipline, academic enterprise should continue without
and the government, organization, or company external interference, enforced disclosure
with whom one consults. Rowan and Miskel becomes an important tool to prevent conflicts of
argue that these social organizations generate the interest from bleeding into research activities.
norms that direct individual behavior (15). The results of this study offer some
Although conventional wisdom suggests that important implications for how academic policies
when the faculty consultant serves multiple should be conceived. First, policy development
masters, academic roles and norms are sacrificed and implementation should rest upon data.
for self-interest, the results imply that the Empirical data provides a foundation for the
consultant maintains an allegiance to the norms formulation of effective and enforceable policy.
of teaching, research, and service. Given these The policies developed in this arena span the
criteria, the faculty in this study can be perceived boundaries between the disciplines, funding
as actors within the institution of academic agencies, academic institutions, and private
science, rather than simply as a set of actors who sector companies. Rather than establish
operate within a single organizational entity. guidelines in isolation of one another, policies
This argument is founded on the capacity of could become aligned across these boundaries to
faculty members to interact in a variety of establish both consistency and clarity.
situations that appear to have competing interests Ultimately, compliance becomes evaluated at
and values while they perfect their craft. If both the department and disciplinary levels.
academic science is the institution, the Consistency and clarity across boundaries will
institutionalized roles and norms embedded in permit faculty to make informed choices.
the scientific method become the criteria Second, policymakers should develop clear
consultant-scholars use to make decisions in their guidelines within their institutional and agency
relationships with commercial scientists. sectors. Policies that guide rather than constrain
University faculty have a societal contract faculty behavior could aid faculty understanding
that affords researchers with academic autonomy of specific behaviors that constitute conflict of
in exchange for a commitment to improve social interest. Furthermore, clearly articulated
welfare through teaching, research, and service guidelines should identify the consequences of
(16). The question that drives university conflict individual action so faculty will understand the
of interest policies is whether or not faculty ramifications of their behavior.
fulfill these institutionalized roles without Finally, academic institutions could identify
serving their own self-interest. If they fail to consulting as a component of the faculty reward
fulfill their duties or pursue their own self- structures. Boyer and Lewis suggested that
interest in the course of their academic activities, consulting could become a means for faculty to
critics would argue that they are involved in a involve themselves in both community and
conflict of interest. However, the conflicts that institutional service (1). Consulting activity
138
–––––––––––––––––––––––––––––––––––––––––––––––––––––– Jones & Louis, The Commercialization of Academic Science
could become an element of faculty development 13. Louis, KS, Jones, LM, & Anderson, MS. Bought and
programs that stimulate faculty vitality and, Sold: Academic Ideas in the Marketplace. Paper
ultimately, productivity. presented at the meeting of the Association for the Study
of Higher Education, Miami, FL. (1998, November 6).
14. Louis, KS, Blumenthal, D, Gluck, ME, & Stoto, MA.
Acknowledgements Entrepreneurs in Academe: An Exploration of
This research made possible by a grant funded by the Behaviors Among Life Scientists. Administrative
National Center for Human Genome Research, National
Science Quarterly, 1989; 34, 110-131.
Institutes of Health (Grant Number HGO0724-01),
administered by the University of Minnesota and Harvard 15. Rowan B, Miskel CG. “Institutional Theory and the
University. For inquiries contact: Dr. Lisa M. Jones, Study of Educational Organizations”. In J Murphy, KS
Postsecondary Education Policy Studies Center, University Louis (editors). Handbook of Research on Educational
of Minnesota, 330 Wulling Hall, 86 Pillsbury Drive S.E., Administration, Second Edition. San Francisco (CA):
Minneapolis, MN 55455-0208, 612-625-0773 (voice), 612- Jossey-Bass Publishing; 1999.
624-3377 (fax). 16. Slaughter S, Leslie LL. Academic Capitalism: Politics,
Policies, and the Entrepreneurial University. Baltimore
Bibliography (MD): The Johns Hopkins University Press; 1997.
1. Boyer CM, Lewis DR. Maintaining Faculty Vitality 17. Emanuel EJ, Steiner D. Institutional Conflict of
Through Outside Professional Consulting. In SM Clark, Interest. N Engl J Med 1995; 332 (4): 262-267.
DR Lewis (editors). Faculty Vitality and Institutional 18. Boyer CM, Lewis DR. Maintaining Faculty Vitality
Productivity: Critical Perspectives for Higher Through Outside Professional Consulting. In SM Clark,
Education. 1985. DR Lewis (editors). Faculty Vitality and Institutional
2. Fairweather JS. Faculty Work and Public Trust: Productivity: Critical Perspectives for Higher
Restoring the Value of Teaching and Public Service in Education. 1985.
American Academic Life. Needham Heights (MA):
Allyn and Bacon Publishers; 1996.
3. Patton CV, Marver JD. Paid Consulting by American
Academics. Educational Record 1979: 175-184.
4. Rebne D. Faculty Consulting and Scientific Knowledge:
A Traditional University - Industry Linkage.
Educational Administration Quarterly 1989; 25(4): 338-
357.
5. Teague GV. Faculty Consulting: Do Universities Have
“Control”? Research in Higher Education 1982; 17(2):
179-186.
6. Marsh HE, Dillon KE. Academic Productivity and
Faculty Supplemental Income. Journal of Higher
Education 1980; 51(5): 546-555.
7. Merton RK. The Sociology of Science: Theoretical and
Empirical Investigations. Chicago: University of
Chicago Press; 1973.
8. Mitroff II. Norms and Counternorms in a Select Group
of the Apollo Moon Scientists: A Case Study of the
Ambivalence of Scientists. American Sociological
Review 1974; 39(4): 579-595.
9. Louis KS, Anderson MS. The Changing Context of
Science and University-Industry Relations. In H
Etzkowitz, A Webster, P Healey (editors). Capitalizing
Knowledge: New Intersections of Industry and
Academia. Albany (NY): State University of New York
Press; 1998.
10. Clark BR. Creating Entrepreneurial Universities:
Organizational Pathways of Transformation. United
Kingdom: Pergamon; 1998: 4.
11. Merton, RK. op cit. 1973: 277.
12. Etzkowitz H, Webster A. Entrepreneurial Science: The
Second Academic Revolution. In H Etzkowitz, A
Webster, P Healey (editors). Capitalizing Knowledge:
New Intersections of Industry and Academia. Albany
(NY): State University of New York Press; 1998.

139
5. Understanding Misconduct
Preventing Scientific Misconduct: Insights from “Convicted Offenders”
Mark S. Davis, Justice Research & Advocacy, Inc., USA
Michelle L. Riske, Justice Research & Advocacy, Inc., USA

Keywords: Equity theory, Prevention, Rationalization techniques, Research misconduct, Responsible conduct
of research, Scientific misconduct

The mere seriousness of certain social behaviors implies the need to prevent them. In the case of
conventional crime, for example, survivors of homicide victims or the victims of physical or sexual
assault, when asked what they want most, often will say they wish the incident had never happened.
For them, a successful homicide prosecution does not bring back the lost loved one. A long prison
term for the rapist will not restore the victim to the state she enjoyed prior to the crime. As a result,
we strive to identify and implement various ways of reducing opportunities for both offending and
victimization.
Although the perceived harm in research misconduct may not be as great as in violent crime, its
consequences nevertheless can have disastrous and far-reaching effects. After-the-fact measures such
as the investigation of allegations and the sanctioning of the guilty, while necessary for justice and the
vindication of the moral order, seldom can undo the harm caused by each instance of fabrication,
falsification, plagiarism, or other serious departure from the norms of science. The retraction of a
published paper cannot restore the time wasted by other investigators pursuing pointless lines of
research or by editors and referees reviewing meaningless results. An apology and a signed voluntary
consent agreement by one found guilty of research misconduct does not automatically lift the taint
from the supervisor and colleagues in whose lab the misconduct occurred. And for those who suffer
from life-threatening diseases and consequently hold out hope for a cure, the broken trust of falsified
clinical trials has far more devastating effects. To be sure, the shock waves emanating from a single
incident of research misconduct can create untold collateral damage, including the tarnishing of
reputations of scientists, institutions, and of the enterprise of science itself.
In view of our collective inability to undo the damage and effect restoration to all parties in these
cases, the prevention of research misconduct is a desirable end. The question then becomes, what can
the scientific community do to keep research misconduct from occurring in the first place? The
purpose of this preliminary analysis is to explore largely untapped data sources in order not only to
advance theoretical work in this area, but also to glean information of practical import.
In order to tackle the challenge posed by prevention, we must acknowledge that prevention can
occur at more than one level. Douglas Weed, employing public health’s notions of primary and
secondary prevention, suggests that we first need to know something about etiology, and he argues
that there are causal factors both internal and external to the scientist who engages in research
misconduct (1) . Examples of internal causal factors would include psychological problems, financial
motivations, or perhaps the desire to hurt others. Causes external to the scientist, on the other hand,
Corresponding author: Michelle L. Riske, J.D., M.A., 849 Cleveland Ave., Amherst, OH 44001, 440-988-8455 (voice), 440-
988-8455 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
are factors such as the pressure to publish, engage in misconduct formulate rationalizations
inadequate training or supervision, or the fierce for their behavior? And what form might these
competition for research grants. rationalizations take? Sykes and Matza, in their
In either case, successful prevention requires research on juvenile delinquency, discuss several
that we somehow interrupt one or more processes of what they refer to as “techniques of
that lead to an instance of research misconduct. neutralization” including (3) :
For example, if we knew that individual • Denial of a victim (Who am I really hurting
psychopathology was responsible for research by fudging these data?)
misconduct, we perhaps could administer the • Denial of an injury (What is the harm?)
Minnesota Multi-phasic Personality Inventory • Condemnation of the condemners (They’re
(MMPI), the Narcissistic Personality Inventory,
out to get me.)
the Psychopathy Checklist, or other psychometric
tools to help us screen out applicants who were • Denial of negative intent (I never meant to
predisposed to engaging in unethical research hurt anyone.)
practice. In an effort to address an external cause • Metaphor of the ledger (For most of my time
such as inadequate supervision, we might here in the lab I’ve been a hard-working,
institute regular meetings between lab loyal employee. I’m entitled to a slip or two.
supervisors and their staff members. All in all, I’ve done more good than bad.)
Is it possible that individuals who commit
Objectives research misconduct may employ one or more of
This pilot study focuses on two individual- these techniques in order to justify their conduct?
level explanations for research misconduct. First, The second perspective employed for this
Cressey’s research on embezzlement in financial study was social psychology’s equity theory,
institutions was examined (2). Cressey’s which speaks to perceived fairness in dyadic
subjects, who largely perceived themselves to be relationships (4). Equity theory is exemplified in
respectable people, had three characteristics in the common phrases “You scratch my back and
common: I’ll scratch yours” and “One good turn deserves
1. A non-shareable financial problem, for another.” Social beings have come to expect
example, one the individual could not reciprocity when dealing with others. If people
discuss without suffering humiliation; perceive they are getting less from a relationship
2. An awareness the problem could be solved than they are given, they may suffer distress. It
by violating the position of financial trust; is common, then, for the ostensibly exploited
person to take measures to relieve this distress
and and restore a sense of equity. In the case of
3. Suitable rationalizations for the embezzle- research misconduct, scientists may be more
ment of funds to resolve their self-concep- likely to engage in misconduct if they believe
tion as a trusted person. they were deprived of what was rightfully theirs,
Applying Cressey’s work to scientific such as the co-authorship on a publication or a
researchers, is it possible that some have non- coveted promotion. Accordingly, individuals
shareable problems, not necessarily financially- may engage in scientific misconduct as a form of
based, which motivate them to engage in retaliation against a coworker or supervisor if
research misconduct? The possibilities could they believe that they have been slighted or
include the inability to produce replicable work exploited.
under pressure, a perceived lack of talent for
research, or personal problems such as marital or Design
emotional difficulties. For example, William Two sources of data were gathered for this study.
Summerlin, the protagonist in one of the best- The first was information from the case files of
known cases of research misconduct, intimated individuals against whom a finding of scientific
that he had been under a lot of pressure from the misconduct was made by the Office of Research
head of the lab to produce results. Could the Integrity (ORI). A standard data collection form
inability to withstand this sort of pressure was used to record data including the institution,
constitute a non-shareable problem? type of alleged misconduct, information from the
In addition to possibly having such non- respondent, response of the institution, and
shareable problems, how do researchers who finding by the ORI. A member of the research
144
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Davis & Riske, Preventing Scientific Misconduct
team read each case file and wrote narrative subjects with the case file reviews. Upon
responses to the items on the data collection form completion of the interviews, the subject list was
summarizing information primarily pulled from given to ORI. Both data collection instruments
the investigative reports by the universities and were approved by an Institutional Review Board
from the investigative reports of ORI and its and by the U.S. Department of Health and
predecessors. These narrative responses were Human Services, Office for Protection from
analyzed for this part of the study. A total of 21 Research Risks.
case files were reviewed for the initial pilot
study. These case files included 16 cases Methods of Analysis
reviewed as part of a pretest, as well 5 additional Because theoretical work on scientific
cases that included cases closed prior to the misconduct is relatively meager, we chose to use
formation of the ORI, i.e., these cases were a qualitative approach borrowed from
handled by the Office of Scientific Integrity phenomenological psychology. Rather than first
(OSI), ORI’s predecessor. searching for evidence of specific theories or
The second source of data consists of propositions, the investigator examines the data
interviews with scientists against whom a finding more for “explication” than explanation (5). This
of scientific misconduct was made by the ORI. results in the listing and preliminary grouping of
Subjects who were included in the first nine case terms or phrases revelatory of, in this case,
files used as part of the pretest comprised the etiology. As a check against possible bias created
sample for this portion of the data collection by prior knowledge or other factors, the analyst
process. Because some scientists approached extracts exact phrases rather than interpreted
could not be located or were unwilling to concepts. Another analyst approaches the data in
participate in the interviews, only three out of the the same way, identifying exact wording to
nine contacted were interviewed. It is possible convey possible sources of causation. The
that the experience of having been accused and second step involves the two analysts coming
found guilty of research misconduct was so together to compare and reconcile their lists. In
unpleasant that some subjects have little interest the third step, the analysts group the phrases into
in dredging up the past. One scientist who common themes or constructs. Finally, the
declined to participate in the study summed up constructs are examined to see if they relate back
his feelings in an e-mail to the senior author: to the selected theoretical approaches in order to
“I am very sorry to disappoint you but after help us interpret and discuss the relevance of
more then ten years I have no inclination to these constructs or central themes in explaining
discuss this issue with anybody. With my very the etiology of research misconduct. For
poor English I found it useless to talk about example, in looking at Cressey’s notion of the
the inquisition. I have no idea what is a (sic)
subject and goal of your research, but I wish
non-shareable problem (6), the analyst would
you a (sic) success in your work in the name of group together those extracted phrases suggesting
justice, science and humanity.” such themes as psychological issues, marital
One of the interviewees summed up his feelings difficulties, financial pressure, lack of
more bluntly when thanked for his time: knowledge, difficulty with expectations of a
“The time is not the problem; it’s the pain of supervisor, lack of supervision, or other problems
having to relive this crap.” an individual might reasonably be uncomfortable
The researchers signed a confidentiality sharing with others.
agreement with ORI to protect sensitive case file Data obtained from the case file reviews and
information. The researchers also took additional from the interviews eventually will be content
steps to ensure confidentiality during the data analyzed using the QSR-NUDIST software.
collection process, by excluding the subjects’ Content analysis is a means of systematically
name and case file number from the data analyzing textual information to find recurring
collection instruments. Subjects were identified themes, issues, and motifs, which can then be
by the assignment of a subject number. To match isolated, counted, and interpreted (7, 8) . If the
files with subjects being interviewed, a list appropriate statistical criteria are met, the data
including the subject name, institution, ORI case will also be analyzed to examine relationships
number, and subject number was created. The among variables in order to assess, for example,
information was only used to link interview if a certain type of misconduct or rank is

145
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
associated with the existence of a non-shareable category included 8 out of the 21 subjects).
problem. Senior Faculty included professors, associate
professors, and directors/heads of departments,
The Sample institutions or clinics. Junior Faculty is defined
The data collected was part of a pilot study to test as assistant professors, postdoctoral students,
the efficacy of the data collection instruments research fellows and residents. Other
developed, which were then used as part of a researchers, including research associates,
larger study examining all individuals against predoctoral students, and administrative
whom a finding of scientific misconduct was assistants, made up the remaining positions (5
made by the ORI as of December 2000. A total out of 21). It should be noted that tenure status
of 21 case files were reviewed for the pilot study. could not be gleaned from the case files.
Many of the respondents held academic positions With respect to the types of research
as either Senior Faculty or Junior Faculty (each misconduct committed by these 21 respondents,
38% of the cases were for plagiarism, 19% were
for fabrication, and 19% were for falsification.
10 Fabrication/falsification made up 14% of the
8 8 cases, and the remaining 10% were for a
Number of Subjects

8 combination of falsification, fabrication, and

plagiarism.
6
5
Results
4
Data from the case files reviewed were analyzed
using the qualitative phenomenological
2 approach.

Senior Faculty Junior Faculty

Etiology
Other
Researchers
The systematic search for possible etiological
Type of Position
factors related to our two theoretical
Fig. 1. Researcher's Academic Position perspectives yielded data in support of both
theories.
Phrases or
Combination elements
10% Fabrication extracted from
19% the case files
Fabrication/ showed evidence
Falsification of non-shareable
14% problems such as
publish-or-perish
pressure, lack of
knowledge or
experience,
difficulty with
supervisor’s
Falsification expectations/lack
19% of supervision,
and personal
problems. These
phrases were
usually extracted
from information
Plagiarism contained in the
38% University
investigative
Fig. 2. Classification of Research Misconduct reports or the
146
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Davis & Riske, Preventing Scientific Misconduct

Job Pressure Lack of Subject Personal Problems Problems with

Matter Knowledge Supervision
Enormous pressure to Understanding of grant Personal insecurity Could not fully satisfy
produce/ Pressure to application process/First the expectations of the
produce by supervisor proposal supervisor/ If supervisor
had more realistic
expectations this incident
might never had occurred
Time factors - short Different interpretation of Personal/Family Supervisor was
deadlines/ Short cut to the normal modes of difficulties demanding in research
save time responsible authorship results
Pressure to keep the Understanding of the Medical illness Lacked proper scientific
system working principles of attribution guidance from supervisor/
in review articles Unsupervised
Insecure position Not able to handle Under personal pressure
position/ Saddled with from supervisor to
responsibilities which in publish data in order to
hindsight were out of secure a research
proportion to subject’s position.
training and experience
Isolated laboratory with Never trained in Negligent oversight/
few peers to discuss appropriate record Deficiencies in oversight/
situation or possible keeping Supervisor’s oversight
career alternatives was inadequate
Difficult job situation/
Stressful job situation
Table 1. Etiology - Non-shareable problem
investigative reports from ORI; therefore, the relationship, some of the phrases taken from the
information is hearsay, as the actual statements case files suggest possible motivation by the
made by the respondent or other interested subjects that could indicate retaliatory conduct in
parties were usually not contained within the case response to perceived exploitation. For example,
files. some of the subjects said that they falsified data
Information obtained from the interviews in order to annoy colleagues or that they were not
also provided evidence in support of a non- recognized for their scientific expertise. Other
shareable problem by the respondent, which may subjects discussed competition in relation to
have contributed to his misconduct. For positions within the university or institution and
example, one interviewee stated: competition for submitting papers for
“How am I going to get a position where I don’t publication.
have to worry every 2-3 years about if I don’t
get my grant I’m gonna be out on the street. Implications for Prevention
This feeling of being a second, kind of a second If we look at the preliminary evidence for our
class citizen. Um, the pressures to produce
papers. And, you know, it was, I knew I was
theoretical questions, we can infer some tentative
doing something wrong, but I reached a point implications for prevention. Information
where I didn’t care.” pertaining to lack of proper supervision or
The data also contained summarized statements training suggests that it might be prudent for
from respondents indicating rationalization universities to implement better procedures and
techniques of denial of an injury, condemnation guidelines for supervisors with respect to
of condemners, and denial of negative intent. employee oversight and monitoring
responsibilities. We found some support that
Although information extracted from the case periodic reviews or audits of research notebooks,
files did not definitively point to instances where as well as the original data collected for all
the subject engaged in conduct in order to restore experiments, by the supervisor may help to
a perceived loss of equity in a dyadic reduce research misconduct. Ensuring that
147
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Denial of an Injury Condemnation of the Denial of Negative Intent

Condemners
No harm done because the Subject had opposite and Fabricated sampling times were
experiments were preliminary, competing opinions to research preliminary and never intended to
unimportant, and had not been performed by colleagues of the be published
published complainant
Worked on several of the articles Allegations by complainant were Going to tell supervisor the truth
which were used as references for an attempt to “get rid of” the after the subject had a chance to
the proposal and therefore subject from the University obtain valid counts, but the
permitted to incorporate these subject didn’t have the chance
materials into the proposal
If there was some faulty reporting
of findings, that it was minimal
since it was not the central issue
of the study
Table 2. Etiology - Neutralization Techniques
employees are properly trained on all traced back.”
experimental techniques prior to performing such So upon his returning to the lab from an extended
experiments could also help reduce the trip:
researcher’s lack of knowledge on the subject “...basically they sat me down and confronted
matter, as well as apprehension about me with the fact that these data sets don’t fit.
And, it was a situation of, uh, what do you say
acknowledging that as a problem. Similarly, if you’re caught red-handed? You know all the
discussing the serious ramifications of research original data was there. It was very easy for
misconduct can also discourage some of the them to go back to the original sets and see
denial its perpetrators use to rationalize their that there were discrepancies.”
actions with such conduct; for example, that This same interviewee briefly contemplated
there indeed is harm associated with these actions trying to cover up the misconduct, but again
that affects a variety of actors and institutions, realized:
including, most importantly, the patient “...it was truly a situation where the record
population. keeping system that I had set up was such that
The three interviews conducted to date have there was no way I could possibly go back
also provided some insights for prevention. One through all the computer files and alter those.
subject credited the careful handling of data for There was, you know, everything, the techs had
his own demise: always printed out paper copies, so there was
“...when the technician did the, you know, do shelves of three ring binders with all the data.
the random stuff, yes, there would be a copy It was a situation of, it can’t be done.”
on the computer, but he would also print out One interviewee felt that training might help
the data, you know, a paper copy and put that prevent some research misconduct:
into their books. So, it was, you know, like, it “I think that there should be more training,
was also like a guarantee that I would study in just the basics of the scientific method
eventually be found out and that it could all be and, you know, what is appropriate, you know,

Evidence of possible motivation to retaliate Evidence of possible motivation to exploit

Made up data to annoy a colleague Future dependent on rapid success in the laboratory
Some friction between subject and others in the lab Laboratory placed too much emphasis on short-term
productivity
Bitter relationship between subject and supervisor Competitive pressure for tenure-track positions
Failed to make changes because upset with others Insecure position
Attempt to get rid of subject
Personal animosity against the subject/Prejudice
against the subject
Table 3. Etiology - Equity Theory

148
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Davis & Riske, Preventing Scientific Misconduct
what is not appropriate in terms of experimental 3. Sykes, G, Matza, D. Techniques of neutralization: A
methodology, in terms of statistics, in terms of, theory of delinquency. American Sociological Review
if you’re going to discard data, you know, what 1957; 22: 664-670.
are the other experimental reasons for 4. Walster, E, Walster, GW. Equity: theory and research.
discarding the data? For example, oh yeah, I Boston (MA): Allyn & Bacon; 1978.
had a sudden sneeze and I sneezed and I 5. von Eckartsberg, R. Existential-phenomenological
botched test tubes or I knocked over this research. In von Eckartsberg, R, editor.
particular test tube, or I tested this particular Phenomenological inquiry in psychology: Existential
agent and found that, oh my gosh, I actually and transpersonal dimensions. New York: Plenum Press;
added 10 times the amount of a particular 1996, p. 3-61.
component, you know, those are valid reasons 6. Cressey, D. Other people’s money: The social
for discarding data. You know, I don’t think psychology of embezzlement. New York: The Free
there’s enough emphasis placed on teaching Press; 1953.
people the proper scientific method.” 7. Denzin, NK, Lincoln, YS. Part I: Method of collecting
Another subject offered what he referred to as an and analyzing empirical materials. In Denzin, NK,
Lincoln, YS, editors. Collecting and interpreting
“easy” solution to the problem of fabrication and
qualitative methods. Thousand Oaks (CA): Sage; 1998,
falsification: p. 35-45.
“What you do, is you have, uh, open laboratory 8. Maxfield, MG, Babbie, E. Research methods for
meetings where everyone in the laboratory criminal justice and criminology. 2nd ed. Belmont (CA):
knows what everyone else is doing. Uh, you West/Wadsworth Publishing Co; 1998.
say you did an experiment that took a hundred
rats, but only five rats came into the, into the
lab, it’s pretty clear that you didn’t do a hundred
rats. Uh, if you’re not there doing the work,
uh, that people think you’re doing or know that
you’re supposed to be doing, uh, so I think, uh,
open laboratories, with regular, uh,
presentations of data prevent that.”

Conclusions
We used a qualitative approach to explore
selected aspects of individual-level etiology of
research misconduct. These preliminary data
offer some tentative support for our theoretical
perspectives. More definitive conclusions will
have to await the collection and analysis of the
data from the larger study.
This research-in-progress also offers support
for certain forms of prevention. These
suggestions, rather than the product of well-
meaning, but less-than-well-informed
commentators, come from those most intimately
involved in actual cases. Returning to the
analogy of crime, learning from those who have
engaged in research misconduct is not unlike
debriefing convicted burglars on what would
have kept them from choosing a particular
dwelling as a target. Who should know better
than those who have done it?

Bibliography
1. Weed, DL. Preventing scientific misconduct. American J
Public Health 1998; 88(1): 125-129.
2. Cressey, D. Other people’s money: The social
psychology of embezzlement. New York: The Free
Press; 1953.

149
The Relative Efficiency of Research Misconduct Investigations Involving
Personal Injury vs. Injury to the Scientific Record
Andrew J. Hogan, Department of Medicine, Michigan State University , USA

Ronald J. Patterson, Department of Microbiology and Molecular Genetics, Michigan State

University, USA

Robert L. Sprague, Department of Kinesiology , University of Illinois-Urbana/Champaign,

USA

Keywords: Efficiency, Research misconduct, Type of injury

Research misconduct investigations conducted by universities and other research institutions are
sometimes highly contentious affairs whose findings are disputed both internally and externally. The
central question of the research reported in this paper is whether certain features of the typical internal
research misconduct investigation contribute to the likelihood of closure or to continued contention.
Most research misconduct investigations undertaken in institutions that receive Federal research
contracts and grants follow the investigational model proposed by the National Institutes of Health
(NIH), described here as the tribunal model. In civil law, similar types of disputes (civil fraud,
misappropriation of property, etc.) are dealt with in adversarial proceedings. One measure of the
efficiency of the typical model for conducting a research misconduct investigation is to determine
how often that model produces a definitive finding, or alternatively how often it leads to further
proceedings.
The objective of this study was to test whether the presence of personal injury associated with a
research misconduct allegation influences the likelihood of a post-investigation proceeding (lawsuit,
grievance, legislative hearing, administrative inquiry, etc.), in the context of the use of the tribunal
model of investigation. We hypothesized that the standard tribunal model, which was designed
principally to protect the integrity of the scientific record, might not be very efficient in addressing
misconduct allegations in which a personal injury was the central feature.

Materials and Methods

Data. Cases were identified in the files of Dr. Robert Sprague of the University of Illinois-Urbana/
Champaign, which contain 1,100 references on the 231 research misconduct cases (hereafter referred
to as the “Sprague files”). The Sprague files consist primarily of copies of news stories in scientific
journals, such as Science and Nature, or academic trade journals, such as the Chronicle of Higher
Education and Lingua Franca.
Sixty-three cases were identified as having adequate documentation of alleged misconduct
Corresponding author: Andrew J. Hogan, Ph.D., Health Services Research Division, Department of Medicine, Michigan
State University, B210 Clinical Center, East Lansing, MI 48824, 517-353-4361 (voice), 517-432-1326 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
involving either a personal injury or an injury to files. The abstractor (Hogan or Patterson)
the scientific record. A personal injury case was identified the type of misconduct alleged
one in which a person directly involved in the (fabrication/falsification, misappropriation of
misconduct allegation identified some kind of intellectual property, other serious deviations,
personal loss, usually misappropriation of retaliation, or other). The abstractor then
intellectual property — plagiarism or the determined the nature of injury based on whether
unauthorized use of confidential information there was an injured party known to the
from grants or articles under peer review. A individual alleged to have committed
scientific record case was one involving some misconduct; if so, the case was classified as one
form of contamination of the scientific record. involving personal injury, otherwise as injury to
Scientific record cases usually involved the scientific record. Next the abstractor coded
falsification/fabrication, but sometimes involved for the type of institutional investigation (tribunal
misappropriation of the intellectual property of or adversarial), based principally on whether the
non-parties to the allegation. complainant was a witness or a prosecutor.
Post-investigation proceedings included The abstractor then determined whether there
grievances filed within the institutions, lawsuits, were other proceedings consequent to the
complaints to regulatory or funding agencies, and institutional research misconduct investigation,
requests to legislative or administrative bodies. A such as:
post-investigation proceeding was classified as a • Internal grievances, discrimination com-
due process case if one or more of the parties plaints, etc.
raised due process issues (hearing notification, • Lawsuits, complaints/appeals to administra-
right to call or cross-examine witnesses, impartial tive agencies, complaints/appeals to legisla-
decision-makers, etc.) related to the research
tive bodies.
misconduct investigation.
In those cases where there was some sort of post-
In the tribunal model of a research
investigation proceeding, the abstractor
misconduct investigation, an individual files an
determined whether due process issues were
allegation with an institution, and the institution
raised.
forms a panel to investigate the allegation. The
Finally, the abstractor examined each
panel is responsible to gather evidence, call and
document regarding the role of the institutional
examine witnesses, and make a finding; in
legal counsel as being supportive, neutral, or
common parlance, the tribunal is prosecutor,
obstructive of the procedural fairness of the
judge and jury. The standard NIH-tribunal model
institutional investigation. The abstractor looked
often attenuates some due process rights
for any references to the role of institutional legal
commonly found in adversarial proceedings, in
counsel regarding the selecting or preparing
particular rights to call or cross-examine
witnesses, selecting or preparing panelists,
witnesses and to present evidence. Current NIH
selecting or preparing administrators, handling
policy suggests that the complainant in such an
administrative problems/complaints, issues of
investigation be treated as a witness, rather than
attorney-client privilege, providing or
as a party.
withholding information, applying legal
In an adversarial proceeding, one party
indemnification, deliberating or making findings,
(complainant) accuses the other party
the preparing or editing of reports, the protecting
(respondent) of misconduct. The parties gather
of parties’ due process rights.
and present evidence, call and examine and
To assure the reliability of the abstraction
cross-examine witnesses. The institution provides
process, the first 20 cases were reviewed by both
an adjudicator to process the allegation, hold
abstractors to establish interrater reliability using
hearings and render a decision. We were able to
a data collection tool. Review of the reliability of
identify no unambiguous cases in which the
the initial cases indicated a 94% agreement on
adversarial model was employed in a research
which documents were relevant to each case, a
misconduct investigation.
70% agreement regarding the type of
Data Collection and Reliability. We
misconduct, and a 91% agreement on whether
reviewed 221 documents related to the 63
the injury was personal or to the scientific record.
identified cases. For each document, a form was
There was a 60% agreement on which documents
completed (see Appendix A) identifying the case
indicated the type of institutional investigation,
name and the document number in the Sprague
152
––––––––––––––––––––––––––––––––––––––––– Hogan et al., The Relative Efficiency of Research Misconduct Investigations
but 100% agreement on the type of institutional In the subset of cases where due process
investigation. There was also 100% agreement issues were raised, any controversies regarding
regarding the existence of post-investigation the role of the institutional attorney in the
proceedings. The reasons for the discrepancies in research misconduct case tended to increase the
the classification of misconduct allegations were likelihood of a post-investigation proceeding by
discussed and resolved before finishing the more than six-fold (see Table 2). However, this
abstraction of the remaining cases. result was only marginally statistically
significant.
Results
No unambiguous cases where the original
research misconduct investigation was Parameter Odds Ratio 95% Bounds
administered using the adversarial model were Upper Lower
found. All of the results related to research Personal Injury 3.39** 11.28 1.028
misconduct investigations which were conducted Attorney
6.50* 46.16 0.92
under the standard tribunal model. controversy
Of the 63 cases described in the 221 Grant Context 0.35 1.88 0.07
documents reviewed, 41% of cases resulted in a Table 2. Logistic Regression Analysis: Likelihood of Post-
post-investigation proceeding, and 69% of these Investigation Proceeding Involving Due Process.
involved a due process issue. Of the 63 cases, n=63 / ** = p < 0.05 / * = p < 0.10
41% of cases involved personal injury, and 70%
of personal injury cases resulted in a Conclusions
post-investigation proceeding. Of the personal Because we were able to identify only two
injury cases resulting in a post-investigation ambiguous cases of research misconduct
proceeding, 61% of these proceedings involved a investigations possibly employing an adversarial
due process issue. model, we were not able to determine whether
Ten percent of the 63 cases involved some the adversarial model would result in fewer
controversy regarding the role of the institutional post-investigation proceedings than the tribunal
attorney. Although we looked for instances where model arising out of misconduct investigations
the role of the institutional attorney was involving personal injury.
supportive of procedural fairness, only negative Under the standard tribunal approach to
statements appeared in the literature examined. research misconduct investigations, cases
Twenty-one percent of cases arose in the context involving personal injury are much more likely
of a funded grant. to produce a post-investigation proceeding. We
Multivariate logistic regression analysis was speculate that the tribunal approach frustrates the
performed to determine the likelihood of ability of personally injured complainants to seek
post-investigation proceedings. The results are redress. From the lofty perspective of protecting
presented in Table 1. Personal injury cases are at the integrity of the scientific record, personal
least 10 times more likely to result in a post- injury cases may often appear trivial or
investigation proceeding than cases involving unimportant and clouded by interpersonal
injury to the scientific record. When allegations bickering that borders on the unprofessional.
are made in the context of a funded grant, the Very often personal injury cases involved
likelihood of a post-investigation proceeding is intellectual misappropriation disputes between
reduced, although this effect is only marginally students or junior faculty and senior faculty
statistically significant. members. In such cases, the administrators and
the members of the tribunal conducting the
Parameter Odds Ratio 95% Bounds
investigation tend to be more the peers of the
Upper Lo wer respondent than the complainant. Complainants,
Personal injury 10.34** 36.46 2.94 rightly or wrongly, often believe that the
Attorney investigation is biased toward the respondent and
3.71 33.39 0.41
controversy that the tribunal procedures prevent them from
Grant context 0.22* 1.12 0.04 making the most effective cases against the
Table 1. Logistic Regression Analysis: respondent.
Likelihood of Post-Investigation Proceeding. ORI’s recent policy statement about treating
n=63, ** = p < 0.05; * = p < 0.10 whistleblowers as witnesses will probably
153
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
increase the likelihood of a post-investigation of attorney-client communications as well as
proceeding by giving complainants even less from lack of incidents to report.
standing than they previously held. Caveats. Most reports of research
In some cases the external funder offers a misconduct are from news stories in scientific or
post-investigation appeals process including a trade magazines (Science, Nature, Chronicle of
full due process hearing, for example, the Higher Education). Reliance on these sources
Departmental Appeals Board of the Department could introduce a possible reporting bias, since
of Health and Human Services. The existence of only the most disputatious cases would be
this appeal mechanism may alter the conduct of considered news worthy. This reporting bias
the original investigation, leading to fewer could significantly affect the prevalence data
post-investigation proceedings. The existence of presented earlier, but probably would not have a
an external appeal mechanism may discourage major effect on the results of the multivariate
some institutions that might be tempted to bias a analysis.
research misconduct investigation toward an NIH/ORI reports on the outcomes of research
outcome most favorable to the institution’s misconduct investigations were also a major
reputation or financial interests; the possibility of source of cases. NIH/ORI reports also contain
disclosure and/or reversal at an appeals hearing relatively few plagiarism/ownership cases, which
could act as a check on such institutional might tend to underestimate the number of
behavior. personal injury cases.
Institutional attorneys may face conflicts of Some observers believe that the handling of
interest when fair treatment of the parties to an research misconduct cases has improved over
investigation is not perceived to be in the time. The results of this study found a slight and
institution’s interest. Legal representation of an statistically insignificant temporal decline in the
organization presents many potential ethical number of cases resulting in post-investigation
pitfalls for attorneys, especially when conflicts proceedings. However, this decline was
arise within an organization, as is the case when confounded by a concurrent decline in the
a university must investigate a research number of cases reported over time. Because the
misconduct allegation against a faculty member cases presented here were identified from the
or student. scientific news literature, this latter decline could
While most judges are attorneys, most be a function of either fewer cases (better
attorneys are not judges and most attorneys are management) or less reporting (declining
trained to act as advocates for their clients. Some newsworthiness) or both. A separate study based
institutional attorneys may see their roles as on a fixed baseline of research misconduct
advocates for procedural fairness, but they also allegations in the institutions in which they arose
understand that a finding of misconduct can carry has been proposed to disentangle these
heavy financial and reputational consequences confounded effects.
for the university as well as the individual
respondent.
Moreover, any of the parties to a misconduct
investigation could become a potential litigant
against the university because of decisions made
during the case by university administrators. Case References (Reference Identifier in
Therefore there may be a strong tendency to act Sprague Database, citation)
as legal advisor to university administrators as
opposed to advocates for a fair and impartial 5310. Marshall E. EPA faults classic lead poisoning study.
investigation. Science 1983; 222: 906-7.
In this research, it is difficult to determine 6411. Weissman v. Freeman. D. New York 1989 Feb 23;
whether controversial actions by institutional Nos. 225, 353.
attorneys was a cause or consequence of post- 6438 . Investigations, S. o. O. a. Fraud in NIH grant
investigation proceedings, since the timelines programs. 1988.
necessary to distinguish cause from effect are 6439 . Crewdson J. Science on the hot seat. Chicago
often missing in the kinds of documents Tribune 1989 Mar 19; Sect. 5, 9, 13.
6483 . Dakins D. Copyright lawsuit illuminates debate over
reviewed. Also the frequency of such reports are academic standards. Diagnostic Imaging 1989 May;
low, but this could arise from the confidentiality 54-60.
154
––––––––––––––––––––––––––––––––––––––––– Hogan et al., The Relative Efficiency of Research Misconduct Investigations
6484. Mervis J. Bitter suit over research work asks “Who 6842. Hamilton DP. NIH misconduct procedures derailed.
deserves the credit?’. The Scientist 1989; 1, 4. Science 1991; 251: 152-3.
6486. Blum D. Yeshiva professor wins copyright lawsuit. 6874. Cordes C. Professor asserts Stanford lied to U.S.
Chronicle of Higher Education 1989 Apr 5; A14. about misconduct probe. Chronicle of Higher
6494. Raymond C. Institutes of Health find that Purdue Education 1991 Mar 13; A21, A24-A25.
professor plagiarized experiments. Chronicle of 6920. Hallum JV, Hadley SW. NIH misconduct procedures:
Higher Education 1989 Jul 12; A2. Effect of new ruling (Letter to the Editor). Science
6496. Marshall E. USGS reports a fraud. Science 1989; 1991; 251: 1296.
244: 1436. 6945. Even misconduct trials should be fair. Nature 1991;
6512. Barinaga M. NIMH assigns blame for tainted studies. 350: 259-260.
Science 1989; 245: 812. 6990. Rooney P. UI researcher steps down after admitting
6519 . Gupta’s defense. Science 1989; 245: 1192. plagiarism. News Gazette 1991 Apr 11; A1, A10.
6521. Erben H. Carelessness, or good faith? Science 1989; 6991. Rooney P. Colleagues see pressure in plagiarism case.
245: 1165-6. News Gazette 1991 Apr 12; A3. Ames, Rusell
6523. Coughlin EK. Geologist accused of misidentifying University of Illinois 19913
fossils calls charges ‘misleading’. Chronicle of 7057. London R. Judge strikes rules on scientific fraud
Higher Education 1989 Sep 13; A9. inquiries. New York Times 1991 May 3; Section 2,
6532. Raymond C. Allegations of plagiarism of scientific 11.
manuscript raise concerns about ‘intellectual theft.’. 7079. Weiss T. Too many scientists who ‘blow the whistle’
Chronicle of Higher Education 1989 Jul 19; A4, A7. end up losing their jobs and careers. Chronicle of
6551. Water T, Oliwenstein L. Science duds. Discover Higher Education 1991 Jun 26: A36.
1990: 42-44. 7090. Monson N. Misconduct in the labs. Health Watch
6565. Zurer P. NIH panel strips research of funding after 1991: 24-33.
plagiarism review. Chemical & Engineering News 7097. Wheeler D. U.S. has barred grants to 6 scientists in
1989; 24-5. past 2 years. Chronicle of Higher Education 1991;
6603. Greenberg DS. Squalor in academe Plagiarist gets A1, A6-A7.
promoted, victim is out of her job. Science & 7216. Wiernik J. Suits charge plagiarism in research. Ann
Government 1990, May 1; 1, 5-7. Gualtieri, C. T. Arbor News 1991 Sep 18; B1- B2.
“Letters.” . 7279. Woolf P. Fraud in science: How much, how serious?
6622. Henderson J. When scientists fake it. American Way 1989 Hastings Center Report 11, No. 5: 9-14.
1990; 56-62, 100-1. 7284. Blakely E, Poling A, et al. Fraud, fakery, and
6623. Heidi S. Weissmann v. Leonard M. Freeman. 684 Fed fudging. : 313-330.
Sup: 1248. 7327. Rennie D. The Cantekin affair. JAMA 1991; 266:
6629. Kaplan, R.J. (1989). Letter. 3333-7.
6636. Scandal upon scandal. Nature 1990; 343: 396. 7328. Winslow R. Earache studies lead to painful dispute.
6661. Gordon G. Lab aide sues U-M, ex boss. Detroit News Wall Street Journal 1991 Dec 18; B1, B3.
1990 May 6: 1A, 9A. 7329. Lead researcher under investigation. Science 1991;
6689 . Culliton BJ. NIH misconduct probes draw legal 254: 1575.
complaints. Science 1990; 249: 240-2. 7486. Ernhart CB, Scarr S. Lead challenge (Letter to the
6728. Hamilton DP. NIH sued over misconduct case. Editor). Science 1992; 255: 783.
Science 1990; 2419: 471. 7501. Moore JA. Cases of misconduct related to speech and
6738. Twedt S. Federal study raps Pitt medical school. The hearing science. Paper for HSS 1991; 494/Psych
Pittsburgh Press 1990 Sep 9: Sect. A:1,7. 493.
6744. Hallum J, Hadley S. NIH Office of Scientific 7551. Roberts L. Misconduct: Caltech’s trial by fire.
Integrity: Policies and procedures (Letter to the Science 1991; 253: 1344-7.
Editor). Science 1990; 249: 1227-8. 7584. Wheeler DL. Scientists begin to question
6745. Weismer G. Orofacial impairment in Parkinson’s confidentially of fraud investigations. Chronicle of
disease (Letter to the Editor). Neurology 1990; 40: Higher Education 1992; A1, A8, A11.
191-3. 7632. Woolf PK. Deception in scientific research.
6747. Hamilton D. Science misconduct legalese thickens. Jurimetrics Journal 1988; 29: 67-95.
Science 1990; 249: 1102. 7673. Adler T. Psychologists criticize study of low lead
6812. Bradshaw RA, Adler AJ, et al. Bridges committee levels. Monitor 1992 Mar; 40-1.
procedures (Letter to the Editor). Science 1990; 250: 7689. Woolf PK. Accountability and responsibility in
611. research. Journal of Business Ethics 1991; 10:
6831. Pinney GW. ‘Whistle-blowers’ are encouraged to 595-600.
come forward at ‘U,’ dean says. Minneapolis Star 7757. Rigert J, Lerner M. ‘U’ researcher could face federal
Tribune 1990 Oct 9: 4B. charge. Star Tribune 1992 Aug 22; 1A, 4A-1A, 5A.
6832. FDA investigating doctor at ‘U’ who led drug study. 8020. Anderson M. Impostors in the temple. New York:
Minneapolis Star Tribune 1990 Dec 9; 4B. Simon & Schuster 1992.

155
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
8070. Rigert J, Lerner M. In courting industry, ‘U’ has 8608 . Rosen JE. Statement of Dr. Jane Rosen. Undated.
invited trouble.” Minneapolis Star Tribune 1992 Dec 8609 . McCutchen CW. “Letter.” 1993 Aug 24.
31; 1A, 14A. 8611 . Rosen JE. “Letter.” 1993 Aug 24.
8186. Scarr S, Ernhart C. Of whistleblowers, investigators, 8612 . Rosen JE. “Letter.” 1993 Aug 12.
and judges. Ethics & Behavior 1993; 3: 73-93. 8627 . Taubes G. Fraud busters. Review of Academic Life
8187. Ernhart CB, Scarr S. On being a whistleblower. 1993; 47-53.
Ethics & Behavior. 1993; 3: 199-206. 8648 . Grassley CE, Cohen WS. “Letter.” 1993 Sep 23.
8197. Minnesota psychiatrist accused of fraud in study. 8791 . Hilts PJ. U.S. scientist says superiors stole credit for
News Gazette 1993 Feb 17: D-2. her vaccine. New York Times 1993 Dec 18; 1010.
8217. Rigert J. Top child psychiatrist at ‘U’ indicted on 8841 . Nimmons D. Sex and the brain. Discover 1994;
charges of fraud. Minneapolis Star Tribune 1993 Feb 64-71.
17; 1A, 11A. 8902 . Butler K, Ohland G. Making the case against
8218. Sweeney P. U professor faces charges on drug study discrimination. LA Weekly 1991 Jan 18.
irregularities. Saint Paul Pioneer Press 1993 Feb 17; 8915 . Anderson D, Attrup L, Axelsen N, Riis P. Scientific
1A, 6A. dishonesty & good scientific practice. Copenhagen:
8259. Cordes C. Research project on ear infection Danish Medical Research Council 1992 .
dramatizes challenge of conflicts. Chronicle of 9097 . Lock S, Wells F, et al., Eds. Fraud and misconduct
Higher Education 1993 Mar 3: A23, A28. in medical research. 1993.
8262. McCuen RH. Institutional oversight to enhance 9185 . Immunologist fined $5,000 in lab misdeed. New York
integrity in research. Journal of Professional Issues Times 1994 Aug 31; A16.
in Engineering 1990; 116: 27-37. 9186 . Final findings of scientific misconduct. NIH Guide
8298. Penn S. Misconduct in the modern university. Thistle for Grants and Contracts 1994; 23: 2.
1992; 1, 7. 9417 . Findings of scientific misconduct. NIH Guide 1995.
8299. Loh P. Vest sued for coverup of scientific 9542 . Hilts PJ. A university and 4 scientists must pay for
misconduct. Thistle 1992; 1, 8-1,11. pilfered work. New York Times 1995 May 19; A20.
8323. Zorza M. “Letter.” 1993 May 10. 9568 . Bauer HH. Ethics in science. [Internet] Discussion of
8419. Cotton P. Harassment linked to fraud, lawsuit alleges. Fraud in Science; 1995 May 1.
JAMA 1992; 267: 783-784. 9603 . Taubes G. Plagiarism suit wins: Experts hope it
8439. Burd S. Shalala’s criticism of U.S. polices on won’t set a trend. Science 1995; 268: 1125.
scientific misconduct troubles some lawmakers and 9614. Herling J. Embarrassment in Hong Kong. Chronicle
watchdog groups. Chronicle of Higher Education of Higher Education 1995 Mar 24; A43-A44.
1993; A31, A34. 9616 . Parrish CJ. The Lumigen Project - or how Wayne
8478. Lueders B. Hunger strike feeds UW controversy. State University went into business and lost its shirt
Isthmus 1993; 5, 7. and a lot more! Newsbriefs Wayne State University
8479. Kerr S. Feds freeze fraud-finders A Wisconsin 1994 Sep 7;. 1-7.
connection? Shepherd Express 1993 Jun 17-24;. 3. 9617 . Shellum B. Chemical conspiracy. Detroit Free Press
8501 . Leo A. Paquette, Ph.D., Ohio State University. NIH 1995 Apr 7; 1E-2E.
Guide for Grants and Contracts 1993; 2-3. 9618 . Wilson M. Judge: WSU chemist stole major
8502 . James H. Freisheim, Ph.D., Medical College of discovery. The Detroit News 1994 Aug 17; 1A, 6A.
Ohio. NIH Guide for Grants and Contracts 1993; 9636 . Lytton H. Frequency of FFP. [Internet] 1995 Jun 11.
1-2. 9642 . Kytle R. HHS release. [Internet] 1995 Jun 23.
8506 . Mark M. Kowalski, M.D., Ph.D., Dana Farber 9649 . Holden C. More on genes and homosexuality.
Cancer Institute and Harvard University. NIH Guide Science 1995; 268: 1571.
for Grants and Contracts 1993; 2. 9654 . Crewdson J. ‘Gay gene’ study comes under fire.
8523 . Burd S. New policy of naming scientists who are Chicago Tribune 1995 Jun 25, 1, 6-1, 7.
found guilty of fraud renews debate on federal role. 9662 . Benditt J. Conduct in science. Science 1995; 268:
Chronicle of Higher Education 1993 Jun 30; A24, 1705-1718.
A25. 9684 . Marshall E. NIH’s “Gay gene” study questioned.
8549 . Final findings of scientific misconduct. NIH Guide Science 1995 Jun 30; 268: 1841.
for Grants and Contracts 1993; 2-3. 9707 . Findings of scientific misconduct. NIH Guide 1195;
8585 . Holland E. Committee finds validity in allegations 24.
against Paquette. On Campus 1993 Aug 19. 9720 . Integrity, O. o. R. Annual Report. Department of
8586 . Lore D. OSU hunts core of plagiarism. Columbus Health and Human Services 1994 Sep.
Dispatch 1993 Aug 23; C1. 9727 . Researcher assessed damages for distribution of a
8587 . Cleverley M. Chemistry professor has student cell line. ORI Newsletter 1994; 3: 1-2.
support. The Lantern 1993 Aug 26. 9739. Burd S. Nursing professor found guilty of scientific
8604 . Polsby M. NIH spasmodic torticollis. Bethesda, misconduct. Chronicle of Higher Education 1995
MD: National Institutes of Health. 1993 Jul 30. July 7; A27.
8606 . Polsby M. “Letter.” 1993 Aug 20. 9757. Bivens LW. Findings of scientific misconduct. Office
8607 . Rosen JE. “Letter.” 1993 Jun 26. of Research Integrity. Notice 1993 June 21.

156
––––––––––––––––––––––––––––––––––––––––– Hogan et al., The Relative Efficiency of Research Misconduct Investigations
9784 . Wilchek M. [Letter to the Editor]. Science 1995; 10772 . Ann Marie Huelskamp, M.H.S. Johns Hopkins
269: 1034. University School of Medicine. ORI Newsletter
9812 . Office of Research Integrity, Annual Report 1994. 1997; 5.
1995. 10789 . Zalewski D. Do corporate dollars strangle scientific
9826 . Findings of scientific misconduct. NIH Guide 1995; research? Lingua Franca 1997; 51-59q.
24, No. 33. 10802. Findings of scientific misconduct. NIH Guide 1997;
9854 . Findings of scientific misconduct. NIH Guide 1995; 26, No. 22.
24, No. 34. 12345 . Mazur A. The experience of universities in
9890 . ORI/AAAS Conference plagiarism and theft of handling allegations of fraud or misconduct in
ideas. 1993. research. Project on Scientific Fraud and
9959 . Case summaries. ORI Newsletter 1996; 4, No. 2: 5. Misconduct: Report on Workshop Two, AAAS 1989.
9963 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 15.
9972 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 13.
9974 . Miller JP. AIDS research scientific fraud. [Internet]
1996 Apr 30.
9975. Findings of scientific misconduct. NIH Guide 1996;
25, No. 13.
9976. Findings of scientific misconduct. NIH Guide 1996;
25, No. 10.
9983 . Jayaraman KS. Court allows fossil fraudster to
return. Nature 1996; 380: 570.
10003 . Cherry M. Black scientist faces inquiry at South
African university. Nature 378: 324.
10079. Findings of scientific misconduct. NIH Guide 1996;
25, No. 22.
10084. Alan L. Landag. ORI Newsletter 1995; 4, No. 1: 5.
10110 . King RT. How a drug firm paid for University
study, then undermined it. Wall Street Journal 1996
Apr 25.
10116 . Jock F. MIT torn by bitter dispute over missile.
Science 1996; 271: 1050-2.
10153 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 22.
10154 . Findings of scientific misconduct. NIH Guide 1993;
25, No. 22.
10163 . Gail L. Daubert. ORI Newsletter 1996; 4: 3-4.
10164 . Jamal Z. Farooqui. ORI Newsletter 1996; 4: 4.
10165 . Andrew Friedman. ORI Newsletter 1996; 4: 4-5.
10166 . Joan Gans. ORI Newsletter 1996; 4: 5.
10212 . Hilts PJ. Dispute on gulf war missile continues to
fester at MIT. New York Times 1996 Feb 24; 11.
10229 . Zinberg D. A cautionary tale. Science 1996; 273:
411.
10260 . Findings of scientific misconduct. NIH Guide 1996;
25, No. 27.
10358 . Garber P. MCG looking at changes admid research
investigation. Augusta Chronicle 1996, October 5.
10463 . Rex D. International recruitment highlights need to
track scientific behavior. Nature 1996, September
12; 383: 107-108.
10605 . Vipin Kumar. ORI Newsletter 1996; 6-7.
10693 . Findings of scientific misconduct. NIH Guide 1997;
26, No. 15. Findings of scientific misconduct. NIH
Guide 1997, May 9; 26, No. 15.
10719 . David F. Eierman. ORI Newsletter 1995; 3.
10724 . Yahya Abdlahi. ORI Newsletter 1996; 5.

157
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Appendix A

DATA COLLECTION SHEET FOR ORI ABSTRACT RESEARCH PROJECTS

CASE NAME________________________ DOCUMENT NO.__________

TYPE OF MISCONDUCT ALLEGED:

(check all that apply)

Fabrication/Falsification ______
Misappropriation of Intellectual Property ______
Other Serious Deviations ______
Retaliation ______
Other:___________________________________

NATURE OF INJURY
(Is there an injured party known to the alleged misconductor?)

Personal Injury _____ Injury to the Scientific Record ______

TYPE OF INSTITUTIONAL INVESTIGATION

(Is the complainant a witness or a prosecutor?)

Tribunal ______ Adversarial______

OTHER PROCEEDINGS CONSEQUENT TO THE INSTITUTIONAL INVESTIGATION

Internal (grievances, discrimination complaints, etc.) ____

If yes, was due process an issue? ____
External:Lawsuits ____
If yes, was due process an issue? ____
Complaints/Appeals to administrative agencies ____
If yes, was due process an issue? ____
Complaints/Appeals to legislative bodies ____
If yes, was due process an issue? ____

ROLE OF INSTITUTIONAL LEGAL COUNSEL

As regards the following, was there any evidence as regards the role of institutional legal counsel as being
(S)upportive, (N)eutral, (O)bstructive or (U)nknown of the procedural fairness of the institutional
investigation? (circle one in each line)

Selection or preparation of witnesses: S N O U

Selection or preparation of panelists: S N O U
Selection or preparation of administrators: S N O U
Handling administrative problems/complaints: S N O U
Issues of attorney-client privilege: S N O U
Providing or withholding information: S N O U
Application of legal indemnification: S N O U
Deliberation or making findings: S N O U
Preparation or editing of reports: S N O U
Protection of parties’ due process rights: S N O U

158
Ethical Evaluation of Misconduct Cases
Doric Little, School of Medicine, University of Hawaii, Honolulu, USA.

Martin D. Rayner, Pacific Biomedical Research Center, University of Hawaii, Honolulu,

USA

Key Words: Ethical evaluation, Misconduct, Scientific divorce

The policies governing the actions of the Ethics Committee at the University of Hawaii were
developed during the late 80’s when the dominant paradigm for Ethics investigations was the
“whistleblower” model. In this model a person of relatively low power in the academic hierarchy
complains of scientific or ethical misconduct perpetrated by a person of higher rank and/or power,
typically within their own academic unit.
For such cases to be handled in an appropriate manner (and to ensure that whistleblowers feel free
to come forward) the confidentiality of the complainant must be carefully protected. Administrative
procedures should minimise the chances that the accused person can use his/her academic power: a)
to have the complaint disregarded without adequate investigation and/or, b) to instigate reprisals
against the whistleblower. However, innocent faculty also need to be protected from frivolous or
malicious complaints. Thus, an initial Inquiry (Phase 1) was required, during which the existence of
the complaint is withheld from the accused, with the accused being informed and interviewed only
after the complainant has convinced the Review Panel that a thorough investigation is justified. At
that point, a full Investigation (Phase 2) is initiated, the accused is informed of the complaint while
his/her lab notebooks, computer files and other pertinent sources of information are immediately
sequestered. The accused then has the opportunity to present detailed rebuttal. If the evidence in
support of this rebuttal seems inadequate, then the committee so reports to the Administration and a
more formal Phase 3 Hearing is set up. It is only after the innocence of the accused has been
reasonably established (typically following the completion of Phase 2) that more difficult issues may
be considered, such as the possibility that the complaint was motivated by envy or by malice.
Furthermore, to conclude that the complaint is malicious requires the committee to assess the
motivations of the accuser at the time the accusation was made. Thus, even if strong suspicions exist,
it is not likely that sufficient evidence will be uncovered to confirm suspicions of malicious intent.
Despite the even-handed principles involved in this approach, the Inquiry Phase of such
investigations is necessarily limited to evidence provided by the complainant. And, more
significantly, both Phase 1 and Phase 2 primarily address the guilt or innocence of the accused. While
we understand that this sharp focus is appropriate in some situations, our experience suggests that this
is not necessarily a “one size fits all” model. This committee has experienced scientific misconduct
cases in which this approach prevented a fair and balanced Inquiry. We suggest that specific
circumstances exist in which policies based on this model may need to be modified to ensure an
appropriately ethical analysis of the complainant’s case.
Corresponding author: Martin D. Rayner, Interim Director, Pacific Biomedical Research Center, 1993 East-West Rd.,
University of Hawaii, Honolulu, HI 96822, 808-956-5184 (voice), 808-956-9574 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Year Complaint Outcome Whistleblower Collaboration $$ issues

protections Breakdown
92 Intel. Prop. Theft Sustained Required No No
92 Plagiarism Sustained Required No No
93 Plagiarism Dismissed Required No No
93 Admin. Miscon. Dismissed No No No
94 Plagiarism Sustained Required No No
95 Authorship Dismissed No Yes No
96 Intel. Prop. Theft Dismissed No Yes No
96 Intel. Prop. Theft Dismissed No Yes No
97 Intel. Prop. Theft Negotiated No Yes No
98 Misapp. of funds Reimbursed Required No Yes
99 Theft/fabrication Dismissed No Yes Yes
99 Intel. Prop. Theft Dismissed No Yes Yes
99 Intel. Prop. Theft Sustained No Yes Yes
99 Sci. Miscond Sustained No Yes Yes
00 Hum Subj. issue Sustained No No No

Table 1. Analysis of cases presented to the University Ethics Committee from 1992 to 2000

Results
Despite the many high-profile cases, nationally, responsible for these changing patterns. First,
which seemed to fit the whistleblower model changes in funding patterns have increased the
during the 80’s and early 90’s, we have noted payoff for collaborations between potentially
significant changes in the nature of the competing laboratories. Second, as scientific
complaints coming before our committee over information has become increasingly regarded as
the last five years (see Table 1). As shown in this potentially marketable intellectual property, it is
Table, six of the nine cases occurring after 1995 inevitable that disputes will arise as to the
involved issues of intellectual property. Before ownership of that property. The stakes are further
this time, however, only one case out of six raised when University Administrators suggest
involved a clear intellectual property dispute. that returns to research units from the marketing
Seven out of the nine cases since 1995, but only of such intellectual property should become a
one out of the six earlier cases, involved significant component of the budgets of
breakdowns in scientific collaborations. academic research units. In apparent response to
Similarly, five out of the nine post-1995 cases these trends, our recent cases have been
involved high financial stakes, whereas none of motivated primarily by disputes over the
the earlier cases seem to have been primarily ownership of potentially valuable intellectual
motivated by financial considerations. Finally, property. These situations are not consistent with
whereas four out of the six early cases required the whistleblower model on which our Ethics
whistleblower protections to protect the identity policies and procedures are based - making them
of a junior complainant, only one complaint out difficult to evaluate. However, these cases cannot
of nine cases since 1995 benefited from such be dismissed as being merely “authorship
protections. Thus, whistleblower protections are disputes” beneath the level of interest of those
still needed, although cases that fit that specific whose duty it is to evaluate true scientific
model are no longer a major part of our misconduct issues, in view of the very high
workload. stakes which may be involved. Finally, we have
seen such cases start at the level of an authorship
Discussion dispute, only to later expand into full-scale
accusations of data fabrication.
Nevertheless, our university’s policies as
Ethics Evaluations in A Changing World
well as the general awareness of the scientific
Two nation-wide trends may well have been
community remain tuned to the whistleblower
160
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Little & Rayner, Ethical Evaluation of Misconduct Cases
model. So, as one might well expect, our cases of such disputes seem to arise from the
continue to be presented in the “approved” breakdown of formerly effective collaborations.
whistleblower format, promising to reveal And, since most collaborations collapse from
significant instances of scientific misconduct. personality conflicts, it is hardly surprising that
If one fails to understand their origins, such such breakdowns lead to disputes over the
cases can be difficult to evaluate. In one such custody of intellectual property. The comparison
instance we were unable even to conclude that a with that other graveyard of failed collaborations,
valid case existed under our restrictive rules for the divorce courts, is inescapable. The level of
Phase 1 Inquiries. What does one do when Phase acrimony over rights to intellectual property
1 of a “denial of authorship” complaint leads to seems fuelled by these underlying personal
the complainant eventually submitting letters issues, just as rights to child custody may
from the accused in which the accused pleads become the focus of a parent’s sense of violation
with the complainant to accept authorship on the in a divorce situation. An Ethics Committee that
paper in question? Should the accused have been must stick its head into a “scientific divorce”
interviewed during Phase 1, in this case, so as to needs to be well aware just how high the
gain additional understanding of the background emotional stakes may have become for the
against which the complaint was made? The individual contestants regardless of the monetary
initial decision that there was no case to pursue, worth of the objective data.
precipitated a seemingly endless series of The committee will need to remember that
requests for external intervention and/or re- not all fights are about money. Some fights are
evaluation of our committee’s policies. We need incomprehensible from any other motive than to
to do better than that. humiliate the opponent. And they will need to
Similarly, other recent cases before our recognise that when it takes at least two people to
committee have seemed to involve inherent bake such a pie, it often takes two to spill it on
conflicts between the superficial appearance and the floor. Of course, the participants in this
the underlying realities of each case. The stage “divorce” may not have behaved equally badly,
now seems set for continuing problems arising, but the party most wronged is not necessarily the
in part, from our evaluative approaches. Perhaps, one who complains the most loudly. This is
significant changes should be proposed in both dangerous territory for an investigative
the published procedures and investigative committee, where the most fundamental
approaches so as to permit effective evaluation of assumptions of the whistleblower model may no
cases that do not fit the whistleblower paradigm. longer be valid.
However, these cases raise arguments for
modifications of our procedures that might, if Formulating a working hypothesis
implemented, remove key protections for more The essence of the issue is this: whereas the
classic whistleblowers. whistleblower model appropriately evaluates the
This seems a potentially dangerous situation validity of the complaint, in a “scientific divorce”
in which it would be all too easy for university it cannot be assumed that the substance of the
faculties and administrations to make serious complaint is valid. Furthermore, it was clear that
mistakes while acting from the highest ethical our case load in Hawaii would not be sufficient
motivations. To address these concerns recent to permit even a minimally rigorous prospective
cases have been re-evaluated to search for study of such cases - which is why we are
potentially generalizable patterns within what presenting our ideas to this meeting. If analysis
had seemed to be “property disputes”. Such a of our experience resonates with the experience
pattern could provide the theoretical grounding of other similar committees, perhaps they will
from which a more systematic approach could be also take up this issue.
developed towards this different class of “Scientific divorces” may need to be
misconduct complaints. evaluated by different procedures. In these cases
Excluding situations involving “priority of one should not focus on the guilt or innocence of
discovery” issues, or situations of outright theft the accused, but rather survey the ethical
(none of which we have yet seen), when two landscape in which the breakdown of
groups feel that they both have valid claims to collaboration occurred. Specifically, it is not
some piece of the same pie this is probably a pie appropriate to assume that the complaint is valid
they baked together. In other words, the majority or that the complainant is not a material
161
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
contributor to the situation under investigation. Next steps
To support this approach, the preliminary The most effective method needs to be explored
instructions given to our Review Panels were by which to merge this “Ethical Landscape
changed. When the initial complaint indicated model” into policies written to protect
that either an intellectual property dispute, or a whistleblowers. We would like to avoid a triaging
breakdown in collaboration, was involved, it was mechanism which would separate cases into, for
suggested that both the complainant and the example: intellectual property cases, misconduct
“accused” needed to be interviewed during Phase cases and “harm/rights” cases with different
1. In other words, it may be impossible to guidelines (as in the the separate courts of our
determine whether or not misconduct is likely to legal system). Instead, we have hoped to find
have occurred unless both parties are some way to treat all our cases from an ethical
interviewed. In a situation of this kind, however, perspective, while at the same time preserving
the committee needs to be aware that the our protections for whistleblowers. We now
complainant will have had time to purge any files believe that ALL cases can be addressed from
that might prove embarrassing, although the this ethical approach in which we do not ask “is
accused may well have been taken by surprise. the accused guilty?” but instead ask “what really
Additionally, even in Phase 2 of the happened?” Once the Panel can answer that
investigation, we suggested that the Review question, then they can consider the extent to
Panel delay considering whether the accused which each participant has behaved in an ethical
might be guilty or innocent of misconduct. First, or unethical manner - and we are ready to ask
they should focus their attention on a different whether any of these behaviors rise to the level of
question: “What happened to create the present scientific misconduct. By contrast, Phase 3 of
conflict?”. However, they should be prepared to the investigation (when this is necessary), should
take as much detailed testimony as necessary to be the point at which standard legal models are
answer that very simple question. Only when the introduced.
committee has reached a clear consensus as to Fortunately, only one small change in our
“what happened”, should they attempt to policies is required to implement this approach.
consider which actions taken by each participant The Review Panel needs the discretion to
might rise to the level of scientific misconduct. interview the accused during Phase 1, should
The danger here is that such open-ended they conclude that this can be carried out without
investigation can get out of hand – the Chair of threat to the complainant. Given that freedom,
the Review Panel may need to remind its the Panel can then adopt either the “standard”
members that focus should be maintained on approach to Phase 1, or the “ethical landscape”
immediately relevant events. approach, as seems most fitting to the case under
These instructions appear to have investigation.
substantially facilitated the appropriate ethical Nevertheless, the open-ended investigational
evaluation of difficult cases. Our Review Panels approach advocated here can lead to unusual
have been models of good committee interactions situations. For example, in one recent case the
where all decisions have been unanimous Committee’s final report to the University
following considerable discussion but without Administration recommended censure not only
significant disputes. This surprising degree of for the accused but also for the complainant
agreement resulted from a comprehensive (whose actions contributed to the wrongdoing),
consensus as to “what really happened” – as well as for a third party who facilitated the
committee members have all felt comfortable situation to his own benefit. To have reported
that “blame”, where blame has been needed, was only on the guilt of the accused would have
fairly assigned. Finally, shared understanding of seemed a violation of our Committee’s ethical
the underlying issues allowed them to make duty in this instance.
some very tough calls in potentially explosive
cases. Even in these hard cases, committees Acknowledgements
appear to have appropriately surveyed each We acknowledge receipt of travel funds from
situation without bias and to have resolved the Alan Teramura, Senior Vice-President for
issue appropriately. Research at the University of Hawaii.

162
Potential Cultural Factors In Scientific Misconduct Allegations
Walter M. Meyer, III, Department of Psychiatry and Behavioral Sciences, The University of
Texas Medical Branch, Galveston, USA
George M. Bernier, Jr., Department of Internal Medicine, The University of Texas Medical
Branch, Galveston, USA

Key words: Authorship disputes, Ethnicity, Gender, Scientific misconduct

Since 1993, The University of Texas Medical Branch has had 16 allegations of scientific misconduct.
They were each examined carefully during an inquiry by a faculty committee and the scientific
integrity officer for evidence of fabrication, falsification or plagiarism. Only one of them was judged
to be scientific misconduct. It involved plagiarism, which was acknowledged by the respondent, and
this case will not be discussed further in this document. The remaining 15 allegations did not reach
the stage of investigation. They involved a variety of other types of complaints: an authorship dispute
in 4 cases, inadequate sharing of data in 3 cases or allegations of questionable research practices in
the remainder. Since many of these disputes involved individuals who were not born in North
America and were raised in different cultural settings, the authors hypothesized that cultural factors
underlie many of these allegations. In order to examine this question, they have done a retrospective
review of the 15 allegations.

Methods
A retrospective review of these 15 allegations was done to detect the possible involvement of gender,
academic status, ethnic factors or cultural concerns. To determine whether any ethnic or cultural
group appeared to be overly represented as complainant or respondent, the cultural/ethnic background
status of the entire faculty, post-doctoral fellows and research technical personnel was compared to
those involved in these allegations.

Results
The 15 complaints involved 29 people; 13 White (10 European descent, 3 Middle Eastern descent),
one African American and 15 Asians (9 Indians and 6 Chinese). See Table I for ethnic distribution of
the complainants and respondents. One of the Indians was involved in two separate instances, once as
a respondent and once as a complainant. All the Asians were born and raised outside of the United
States. Six of the complainants were White (4 European descent, 2 Middle Eastern descent) and 3 of
these were born and raised outside of North America. Seven of the respondents were White (5
European descent, 2 Middle Eastern) and two were born outside of North America. The one African
American individual, born in the United States, was a respondent. Nine Asians (4 Chinese and 5
Indians) were complainants and 7 Asians (2 Chinese and 5 Indians) were respondents.
Corresponding author: Walter J. Meyer III, Department of Psychiatry and Behavioral Sciences, University of Texas Medical
Branch, Galveston, TX, 77555, 409-747-8355 (voice), 409-747-8351 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Complainants
White, White, Asian, Asian, African
Total
US Foreign Indian Chinese American
White US born 1 2 1* 1 0 5
White Foreign born 0 1 0 1 0 2
Asian, Indian 2 0 3* 0 0 5
Asian, Chinese 0 0 0 2 0 2
African American 0 0 1 0 0 1
Total 3 3 5 4 0 15

Table I: Number of complainants and respondents by ethnic group

* One person was a complainant and a respondent
other instance, the complainant referred to the
Three subjects involved in these allegations political leanings of the respondent as they
were technicians, seven were post-doctoral related to their country of origin, i.e., brands of
fellows and the remaining 19 individuals were communism.
faculty. One faculty was involved in two To determine whether any ethnic or cultural
allegations, once as a complainant and once as a group appeared to be overly represented as
respondent. The complainants and the complainant or respondent, the cultural/ethnic
respondents were of very similar ages, mean ages background status of the entire group of
of 45.7 and 44.0 years, respectively. In ten cases, university employees (faculty, bachelor level
the complainants were older than the respondents technicians or post-doctoral fellow) was
and in five they were younger. Ten of the compared to those involved in complaints. All
complainants were of lower rank in the university but one female professor was or had been
than their respective respondents. Only five of employees of the university. Only five of the
the 29 individuals were female (two Whites, two individuals were female (two Whites and three
Indians and one Chinese). These 5 women were Asians). The faculty is 24% female and 17% of
involved in a total of 3 allegations. these allegations involve females.
Six of the allegations involved individuals There is a great difference in the ethnic
from different ethnic groups. The remainder distribution of the total faculty compared to those
involved individuals from the same ethnic or individuals involved in scientific misconduct
cultural background. Of the six disputes allegations. The medical school has a faculty of
involving more than one ethnic group, three 750 individuals (550 White, 39 Hispanic, 24
involved White of European origin and Indians; African American and 136 Asian). Of the 136
two, a White and Chinese; one, a African Asian, at least 55 are from India and 43 are from
American and an Indian. Nine disputes involved China. Table II illustrates the differences in
individuals from the same ethnic group: two ethnic distributions between the faculty, bachelor
involved Chinese; three involved Indians; and level research technicians and post-doctoral
four involved White. Among the disputes fellows at large and those individuals involved in
involving White as both complainant and scientific misconduct disputes. There is a
respondent, one involved both parties being from significant difference between the individuals
the Middle East; one involved both parties born involved in scientific misconduct allegations and
in the USA and of European descent; one the total group of individuals in the same
involved a complainant born in an eastern block category for the faculty (p <.0001 by chi-square),
country and a respondent born in the USA; and the technicians (p <.0001 by chi-square) and the
the last involved a foreign born middle eastern post-doctoral fellows (p <.001 by chi-square).
(Egyptian) complainant and an American born The country of origin was not discerned for the
Israeli respondent. Two of the allegations faculty. But there does seem to be among the
involving Asians referred to deep-seated distrust White individuals an unexpectedly large number
of individuals from similar backgrounds in their
country of origins. In one instance, the of individuals born in the Middle East.
complainant stated that he knew that the
misconduct had occurred because people from Discussion
the village of the respondent were evil. In the In the early 1990’s many universities started

164
––––––––––––––––––––––––––––––––––––– Meyer & Bernier, Potential Cultural Factors in Scientific Misconduct Allegations

White Nat. Am. Hispanic Indian Asian Total

Total Faculty* 73.0 3.2 5.2 0.5 18.1 100
Total Technicians** 56.6 4.6 9.8 0.0 29.0 100
Total Postdoctoral *** 40.0 3.0 4.0 1.0 52.0 100
Involved in Faculty * 52.6 5.3 0.0 0.0 42.1 100
Scientific
Misconduct Technicians** 33.0 0.0 0.0 0.0 67.0 100
Disputes Postdoctoral *** 28.6 0.0 0.0 0.0 71.4 100

Table II: Differences expressed as percent of total in ethnic distributions between the faculty and postdoctoral fellows at
large and those individuals inovlved in scientific misconduct disputes
*Significantly different p <.0001 by Chi Square, **Significantly different p <.0001 by Chi Square
***Significantly different p <.001 by Chi Square

establishing a very formal process to consider were the target of the complaints. The little we
scientific misconduct charges. The initial do know refers only to the respondents who have
definitions were focused on fabrication, been determined to have committed scientific
falsification and plagiarism but did leave an misconduct. We know little about those who
opening for examining ‘other practices that brought the complaint forward because of the
seriously deviate from those that are commonly appropriate concern about damaging the
accepted within the scientific community for whistleblower. Also almost nothing has been
proposing, conducting or reporting research’ (so written about those allegations, which did not
called unusual or questionable practices) (1-3). meet the definition of scientific misconduct as
The allegations or complaints were usually none defined by falsification, fabrication, and
of these; rather they reflected personal disputes plagiarism. One study of authorship disputes
between the complainant and respondent. received at the Ombuds office of Harvard
Questionable research practices were particularly Schools and affiliated hospitals reported that the
difficult to define and often the scientific number of disputes has greatly increased between
integrity officer and/or relevant faculty 1991-2 to 1996-7 (5). Women were involved in
committee were called upon to make a judgment the majority (53%) of the complaints and non-US
of intent. Therefore these disputes were almost citizens were involved in 21% of them (5). The
always impossible to discern with any assurance current study seems to be the only other venture
for fairness. In order to gain insight into these into this area. This study identifies a higher than
types of complaints, a fairly large amount of expected number of individuals who were born,
work has been done nationally to examine the raised and partially educated outside of the
nature of the complaint. In fact, certain types of United States. In addition, the complaints are
complaints such as authorship complaints were often against individuals from the same ethnic
rejected as scientific misconduct. Also the Office background and gender as the complainant. This
of Science and Technology Policy has data is provocative. If substantiated in other
established, but not formally implemented, a universities, it indicates a need to reexamine our
more narrowed definition to exclude questionable education of faculty and post-doctoral fellows
research practices and to include with fabrication, concerning the proper use of the scientific
falsification, and plagiarism only the misconduct complaint process. Also other
inappropriate use of documents which might be mechanisms need to be identified to help settle
seen as part of the review process (4). Even with these misunderstandings among scientific
this narrower definition the complaints about colleagues.
authorship, data ownership and access and There are significant hazards to doing this
questionable or sloppy research practices will type of retrospective review. This type of
continue to plague the university committees and endeavor invites accusations of racism, gender
scientific integrity officers. bias, and other un-American activities, such as
In contrast to open discussion about the racial profiling. In order to get different
nature of the complaints and allegations, almost perspectives on this issue, the authors had the
nothing has been written about the nature of Director of our Affirmative Action Office and a
those who made the complaints or those who member of our Institute of Medical Humanities
165
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
review this manuscript. We are attempting only Bibliography
to describe as a group the complainants and 1. U.S. Public Health Service. Responsibility of PHS
respondents, not to speculate why one group awardee and applicant institutions for dealing with and
rather than another might utilize the scientific reporting possible misconduct in science. 54 Federal
misconduct complaint process to address other Register 151; 42 CRF Part 50 (8 August 1989), 32446-
51.
related issues in the research group setting. One
2. Panel on Scientific Responsibility and the Conduct of
speaker at the recent ORI conference on research Research. Responsible Science: Ensuring the Integrity
(6) suggested that misconduct complaints are of the Research Process, I. Washington DC: National
increasing because of the increased collaborative Academy Press; 1992.
nature of research and increased difficulty in 3. Dresser R. Defining scientific misconduct: the relevance
obtaining funding. Only three of our allegations of mental state. JAMA 1993; 269: 895-97.
involved collaborations outside of the 4. Office of Science and Technology Policy. Federal policy
complainant’s research group. Four of our on research misconduct. 65 Federal Register 235; (6
allegations could be linked to some financial December 2000), 76260-64.
5. Wilcox LJ. Authorship: the coin of the realm, the source
factors but they did not seem to be the main
of complaints. JAMA 1998; 280: 216-7.
issue. Usually the complaint involved very poor 6. Little D, Rayner M. Ethical evaluation of misconduct
communication between the respective parties. cases. Research Conference on Research Integrity.
Some ground rules for working together need to November 18-20, 2000.
be taught as part of the research curriculum. 7. Meyer WJ. Data Management: recording, retention,
access and ownership. In: David TP, editor.
Conclusions Management of Biomedical Research Laboratories: A
The vast majority of complaints did not involve National Conference Proceedings. Tucson, AZ:
scientific misconduct as currently defined. This University of Arizona Press; 1999. p. 173-78.
8. The University of Texas Medical Branch Web Page
retrospective review suggests that cultural accessed at http://www.utmb.edu/avpr/integrity/
concerns may contribute to the complaints to the memo.htm.
scientific integrity office. Proportionally the
Asian group is over represented in the scientific
misconduct complaint process. This report
documents for one university the magnitude of
the apparent influence of cultural differences in
the scientific misconduct complaint process. On
the surface, this retrospective review suggests
that cultural differences account for many of the
authorship and other scientific misconduct
disputes. Since the vast majority of complaints
in this retrospective review did not involve
scientific misconduct as currently defined, we
believe there is a need for an increased
educational effort on the part of the university to
orient faculty, bachelor level research technicians
and post-doctoral fellows on the appropriate use
of the scientific misconduct process and to
develop other mechanisms to help them resolve
conflicts with fellow scientists. Guidelines for
data ownership and management (7), authorship
of grants, and authorship of papers (8) have been
recently established on our campus to aid in this
process.

Acknowledgment
The authors wish to express their gratitude to
Deborah Reynolds for her help in preparing this
manuscript.

166
Whistleblowers in Environmental Science, Prevention of Suppression Bias,
and the Need for a Code of Protection*
Elihu Richter, Occupational and Environmental Medicine, Hebrew University School of
Public Health and Community Medicine, Jerusalem, Israel

Colin L. Soskolne, Department of Public Health Sciences, University of Alberta, Edmonton,

Canada

Joseph LaDou, International Center for Occupational Medicine. University of California

School of Medicine, San Francisco, USA

Tamar Berman, Occupational and Environmental Medicine, Hebrew University School of

Public Health and Community Medicine, Jerusalem, Israel

Keywords: Code of protection, Environmental science, Suppression bias, Whistleblowers

Suppression bias is the distortion in the estimate of findings on hazard and risk inimical to special or
national interests, and is well known (1-4). The direct and indirect repercussions of suppression bias
are issues of direct importance not only to environmental scientists and health and safety
professionals, but also to the public itself. These repercussions raise questions as to the adequacy and
degree of protection provided by professional organizations, research institutions, and the legal
system against such suppression bias.
Suppression bias is rooted in the way societies react to troublesome information, as we know
from the tradition of shooting the messenger of bad news. The trial of Socrates served as the classic
case study of the risks to messengers. The jurors of Athens, a city besieged from without and insecure
from within, convicted Socrates and sentenced him to death for corrupting the morals of the youths of
Athens (5-6). Legal scholars have pointed out that Socrates would be convicted by a modern jury for
the same reasons that he was convicted by the jury in Athens: his teachings undermined order,
stability, and state security. For Athenians, there was a Benthamite rationale for putting Socrates to
death: silencing him was necessary to preserve the greatest good for the greatest number in a society
weakened by external wars and internal divisions (7).
Environmental scientists and occupational health and safety professionals measure and report
health risks from exposures to toxic and physical agents so that preventive measures can be put into
* An earlier version of this paper was published in the Int Journ of Occ and Env Health, 7:68-71, 2001, based on a
presentation at the Collegium Ramazzini Annual Meeting in Carpi Italy on October 29 2000. We thank the Councilors and
members of the Committee on Ethics and Philosophy of the International Society for Environmental Epidemiology for
advice and encouragement, the International Student Foundation and the Collegium Ramazzini for financial support, and Dr
Herbert Levine (Bethesda MD), Mr Ron Dror (Tel Aviv) and Dr Richard Laster (Jerusalem) for their helpful insights.

Corresponding author: Elihu D. Richter, Hebrew University-Hadassah School of Public Health and Community Medicine,
POB 12272, Jerusalem, Israel, 972-2-6758147 (voice), 972-2-6784010 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
effect. We define epidemiologic messengers, or settings where protection of human rights is
whistleblowers, as persons who are subjected to weak. However, information is not readily
harassment, lawsuits, ostracism, job loss, loss of available on the degree to which this or other
funding, intimidation, abuse, threats, or even Federal agencies defined institutional
force after reporting such risks, or are prevented responsibilities to protect investigators from
from investigating or reporting risks altogether. external or internal harassment.
In most scientific fields, the rewards go to
investigators who report «positive findings». But The context and content of the problem
in the environmental sciences, the situation is the Martin (8) has listed the five methods of
opposite. In environmental and occupational suppression bias. These are: (a) preventing
medicine, and in epidemiology and related creation of data (b) controlling, (c) blocking,
disciplines, “positive” findings about hazards (d) distorting data, and (e) attacking researchers.
and risks are threatening to powerful interests. This simple list shows that using harassment to
Investigators who study or report these risks are block dissemination of data on hazard and risk
therefore at increased risk for harassment by the and attacking researchers who report such
very nature of their work. findings are only part of a syndrome of
Ultimately, suppression of information about suppression bias, leading to what is known as
hazards and their health risks may itself become lamppost toxicology or epidemiology. Martin and
hazardous to public health. There has not been Deyo have reviewed the driving forces, context
sufficient recognition of the possibility that such and methods of harassment of epidemiologic
pressures may serve to deter investigation or messengers or whistleblowers, and have provided
assessment of health risks from exposures, and case studies (1, 2, 8).
thereby delay or block the implementation of
preventive measures. So far, there have been few The reported distribution of the problem:
systematic efforts to examine the impact of such sentinel episodes
pressures on the direction, content, and work Does suppression bias deter the prompt
output of environmental epidemiologists, detection, reporting and prevention of hazard and
physicians in occupational medicine, and other risk? If so, is this bias systematic, episodic, or
scientists. Nor has there been sufficient attention sporadic and what are its distributions and
as to how to respond to these pressures. determinants? The details of whistleblower
harassment are not frequently publicized (9), but
Methods below we present a list of episodes that have
This paper reviews past reports and summarizes come to light in the past years from reports
work now being carried out by the ISEE gleaned from the professional and lay literature,
Committee on Philosophy and Ethics and the and from our own direct contacts.
Collegium Ramazzini. This work documents
episodes of harassment of environmental Cases of suppression by a governmental
scientists and episodes of responding to requests institution
for assistance from environmental messengers
• Cate Jenkins, an environmental scientist
subject to harassment. We also make
recommendations for future action by with the US EPA, claimed that chemical
governmental organizations, which define industry studies had consciously minimized
standards for research policy. the hazard of dioxin (10-11). She received a
written reprimand for writing down what she
Findings knew about the history of the dioxin incin-
In the 1980’s, the United States Environmental erator regulations (12-13), and was trans-
Protection Agency (EPA) published a document ferred from her position.
which described the hazards unique to • Omar Shafey, an epidemiologist in the
environmental scientists and the forms of Florida State of Health, was forced to leave
harassment to which they may be subject. It his position after publishing an epidemio-
made the point that harassment is most likely
logic report on complaints of acute illness in
directed at younger or less well-known scientists,
employees of government or industry, or residents exposed to drift from aerial spray-
members of the exposed population itself in ing of malathion, used to control the Medfly
168
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science

(14). contamination of the nationwide milk supply

• Desi Mendoza Rivero, a Cuban physician, with organochlorines (24).
was imprisoned after he issued statements
regarding an epidemic of dengue fever (15). Cases of suppression by industry
• Grigory Pasko and Alexander Nikitin, • In the 1940’s, Randolph Byers, the Harvard
government scientists in Eastern Europe, pediatrician, was sued for defamation and
were accused of treason and subjected to damages by the Lead Industries Association
physical abuse after they reported dangers for publishing findings on brain damage
from nuclear waste in Murmansk (16-17). from acute lead poisoning in children from
From newspaper reports, it appears that nibbling paint chips (25-26).
Pasko’s subsequent acquittal was reversed. • Doug Johnson, a safety specialist for
(17) Tatitlek, Chugach, and Chenega Corporation
• Melvin Reuber, a toxicologist at the in Alaska was fired after raising environmen-
Frederick Cancer Research Facility in tal concerns regarding Alyeska’s oil spill
Maryland (which is part of US National response program in Prince William Sound
Cancer Institute) studied links between (27).
pesticides and cancer. As a result of his • Myron Mehlman, a Mobil Oil Corporation
studies, he is one of the world’s leading toxicologist, was fired after advising a Mobil
critics of pesticides. In 1981, he was sub- subsidiary in Japan to stop selling gasoline
jected to an attack on his work and his with hazardous levels of benzene, a known
credibility that shattered his career (18-19). carcinogen (28).
• In the United Kingdom, a Health and Safety • Alexandra De Blas of Australia was threat-
Executive (HSE) memo indicates that ened with a suit for defamation by a mining
several researchers and health and safety company when she attempted to publish a
activists who exposed poor health and safety thesis about environmental impact of its
practices were targeted for special surveil- operations. (29).
lance (20). • Dr Yoram Finkelstein, an Israeli
neurotoxicologist with important publica-
Cases of suppression by an academic tions on organophosphates and lead, is
institution currently the target of a SLAPP (Strategic
• John Coulter, a medical researcher at the Lawsuit against Public Protestors) lawsuit
Institute of Medical and Veterinary Science for libel after writing a medical opinion on
in Adelaide, South Australia was dismissed the health risks from emissions of hexavalent
from his post after releasing a report that chromium, Cd, lead, Ni, and other pollutants
ethylene oxide was mutagenic (21). from an aluminum foundry (30).
• Robert van den Bosch of the University of
California, Charles Lincoln of the University Survey Results
of Texas, and Robert Fleet of Texas A&M At the Annual Conference of the International
University all suffered abuse because of their Society for Environmental Epidemiology (ISEE)
research on the hazards of pesticides (22). held in 1999 in Greece, the Committee on Ethics
and Philosophy distributed a questionnaire to the
• David Kern, an occupational physician and
delegates. Out of 10 individuals who completed
epidemiologist at Brown University Medical the questionnaire, five reported harassment
School, received notice that his tenure would following publication of research findings on
not be renewed and his clinic closed after he health risks from environmental exposures. The
reported numerous cases of interstitial lung following is a brief description of these cases:
disease in nylon flockers at Microfibres (23). • Male MD, age 47, a scientist in a major
• In Israel, Dr Jerome Westin was greylisted Cancer Institute in Italy, experienced ostra-
for any governmental or academic appoint- cism after publishing findings on asbestos
ments after publishing findings on massive exposure in a petroleum refinery and lung
169
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

cancer. outside pressures and be governed by the criteria

• Female MD, MPH, age 60, was threatened used for peer review. In the second case,
with loss of her job after publishing findings Professor Viel from France reported to the Ethics
on TCDD exposure and cancer. and Philosophy Committee being the target of
job threats following publication of papers in the
• Male MPH, PhD., age 53, experienced
British Medical Journal on risks for cancer
ostracism and the threat of job loss after among seashore residents living near nuclear
publishing findings on cancer mortality in power plants. This investigator also reported
Vietnam veterans exposed to Agent Orange. pressures from the nuclear industry to reveal the
• Two Female MD, investigators age 59 and identity of individuals whose health records were
47, experienced both ostracism and confisca- part of an epidemiologic study. The Ethics and
tion of data after publishing findings on Philosophy Committee convened an ad hoc
ethylene oxide exposure and breast cancer. subgroup, under the late Professor John
Goldsmith, one of its founding members, which
Pressures on institutions communicated with Professor Viel, and offered
Deyo et al have reviewed Congressional to provide moral support for the issues raised. In
harassment of the CDC Injury Prevention Unit both the Needleman and Viel cases, the issues of
following its epidemiologic work on impact of concern were resolved, but it is not known
gun control laws on violent deaths (2). whether and to what degree ISEE’s response
played a role. Both Needleman and Viel are
well-known senior investigators who published
Actions to date:
their work in prestigious journals. Their
The International Society for Environmental
situations are exceptions to the rule that most
Epidemiology (ISEE) Committee on Ethics and
whistleblowers do not have the protection of
Philosophy and the Collegium Ramazzini
status and seniority, their findings or warnings
Committee to Protect Whistleblowers are
may not be particularly original, and they may be
working in parallel to provide moral and
prevented from either from publishing their
professional support to whistleblowers (31). The
findings or completing investigations in
ISEE has already developed procedures designed
progress.
to provide an international service of advice,
Through 2001, ISEE has responded to two
referral, and support for environmental
cases, that of Yoram Finkelstein and Omar
whistleblowers which was first presented in a
Shafey, and is working on a third, that of a
special workshop at the ISEE International
Conference in Athens in 1999 (not far from the pathologist sent to prison in Belarus.
site where Socrates was convicted.) The
Collegium Ramazzini is now doing the same, and Discussion
is planning to expand media reporting of The case studies above provide support for the
whistleblower harassment, with particular hypothesis that powerful governmental, military,
attention to occupational medicine professionals economic, and political interests are often the
in developing countries. The aim of both driving forces and the sources of legal and illegal
professional societies is to establish systems for harassment of environmental messengers and, at
monitoring and reporting harassment and abuse times, the institutions they work for. But most of
of whistleblowers, and to offer support and the case reports are from Western countries with
assistance should it be requested. developed research cultures and codes for the
In 1996-97, before ISEE developed these protection of human rights. The high-risk
procedures, it reacted to two situations in which settings for exposure to pressures against
investigators were subject to political pressures environmental scientists are those where research
resulting from the publication of their findings. is most needed, i.e., where exposures and risks
In the case of Dr. Herbert Needleman, ISSE sent are severe, where there are few environmental
a petition signed by many of its members to the scientists, and occupational safety and health is
University of Pittsburgh asking that its review of not properly regulated and enforced by law. The
the validity of his findings on the effects of low risks are increased where legal safeguards for
level lead exposure on intelligence scores, human rights are weak, and where access to a
behavior, and mood status be insulated from free press is blocked.
Yet, data are not readily available to examine
170
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science
the working hypothesis that the exposure settings endangered public health (33).
in which scientists are at greatest risk for threats,
harassment, and legal pressure are those in which Institutional safeguards against harassment in
they are most needed. Africa, Latin America, environmental science
Asia, the Mid-east and Eastern Europe are the Until now, research on ethics in environmental
regions of the world with the worst epidemiology has focused on the obligations of
environmental and occupational health problems, individual researchers to comply with norms of
the fewest environmental scientists, and the truth and not engage in scientific misconduct
weakest safeguards to protect the rights of (34-35). But there has been insufficient
investigators. The situation is probably the worst discussion of the obligations of institutions to
for physicians working in occupational medicine protect their workers and their findings from
who serve remote populations, given their external harassment when their findings are
relatively low status on the professional totem embarrassing, costly, or threatening to powerful
pole. In many of these countries, the situation for interests. Such harassment serves as a deterrent
environmental scientists parallels the situation to investigating and reporting information about
with regard to human rights, and suppression hazards and risks.
bias, like poor working conditions, is accepted as Measures to protect messengers in
part of the normal research environment. It environmental and occupational epidemiology
therefore stands to reason that in these regions, should be required of grant recipients of research
the absence of information on harassment of contracts around the world and should become a
researchers can almost be said to be evidence of worldwide institutional norm.
the effectiveness of suppression bias as a
deterrent to investigation of environmental
Messengers can be wrong
hazards. So far, neither the ISEE nor the
The statements made by epidemiologic
Collegium Ramazzini have received requests for
messengers on the presence of a hazard or risk
help from these settings.
may be right or they may be wrong. We suggest
In the developed countries, we need to ask
that pressures, harassment, and abuse are no
whether a more subtle institutional form of
substitute for access to the peer-review process.
suppression bias could be taking hold. Academic
At the same time, there is the need to be
institutions are entering into strategic business
concerned about pressures on this peer review
alliances, most often with biotechnology and
process by new trends in the academic world to
pharmaceutical firms (32). The close ties
forge alliances between industrial or
between university and business are a frontal
technological interests and the research
assault on the last vestiges of “academic
freedom” of the faculty members. Moreover, the community.
diminishing role of governments in funding
public health research causes academic What Next?
institutions to pursue corporate funding. This Professional societies derive their legitimacy
trend furthers the alliance of university and from their mission in promoting the public good.
business, and increases the likelihood of Investigation and reporting environmental
suppression bias. hazards and their risks are essential to prevent
We suggest that suppression bias and the damage to public welfare. As we noted at the
occurrence of environmental hazards circularly outset, the protection of epidemiologic
reinforce each other. Alibek has pointed out that messengers derives from the primacy of
in the former Soviet Union, suppression of protecting public health. Ironically, Benthamite
information on health hazards to personnel and rationales —stretched somewhat—could have
the environment from activities related to served to acquit Socrates were it to have been
weaponizing bacterial and viral organisms for shown that his teachings were necessary for
bioterrorism led to a scenario in which safety was protection of the greatest good for the greatest
jeopardized over and over again in the name of number, or more fundamentally, for the health
national security. He described a scenario in and welfare of all individuals, in keeping with
which suppression bias resulting from the traditions of the sanctity of preserving individual
harassment of epidemiologic messengers human life.
Organizations concerned with ethics in
171
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
science in recent years rightfully called attention obstruct access to or distort the process of peer
to the need to establish rigid standards for review.
preventing scientific misconduct by individuals. There is a need to recognize a dissonance
The first generation of work on ethics in research between the emphasis of the first generation of
focused on setting standards, procedures and ethics on promotion of research integrity and that
codes of practices which defined responsibilities of the second on prevention of suppression bias.
of individual scientists at all levels, to work Often there is a two-stage scenario in which
according to codes of truth, quality assurance and investigators—or officials in need of a rapid
quality control, precision and accountability estimation of hazard or risk— are first blocked
(36-37). This first generation of work addressed from access to the exposed populations and
issues raised by whistleblower scientists who relevant databases, and then their reports are
drew attention to scientific misconduct in the disqualified because they are incomplete,
laboratories of their superiors. These episodes of imperfect or imprecise. In short, the very criteria
misconduct led to the distortion of findings, used to define the quality of investigation may
failures in quality assurance and quality control, serve as barriers to reporting its substance. This
and lapses in precision and accountability. The situation —in which being precisely wrong is
issue at hand now is standards for preventing considered preferable to being approximately
institutional misconduct. There has been no right—is the classic scenario of delay.
parallel effort of equivalent force to enact One form of harassment of environmental
standards that prevent misconduct by epidemiologists and other investigators is to
institutions—be they the scientist’s employer or subject their databases and records to a battery of
other bodies—which results in harassment of legal subpoenas. If transparency is our norm, it
epidemiologic messengers. is hard to fault such challenges. However, such
We suggest that failure to ensure proper subpoenas pose potential challenges to the
access to independent peer review insulated from privacy of research on individuals, and may serve
internal and external pressures is a form of as a deterrent to their giving permission to use
institutional misconduct. The same statement data on individual exposure and risk. But, in the
applies to failure to provide protection against case of environmental epidemiology and related
legal harassment, such as occurs with SLAPP fields, the ultimate form of harassment is to deny
lawsuits. Therefore, the second generation of the investigator access to databases, so as to
work in ethics and scientific integrity has to deal prevent a complete investigation. In
with a new and different set of problems. These epidemiology, in particular, barriers to accessing
pertain to the need for standards, procedures, and databases on denominators can be particularly
codes of practice that define the responsibilities devastating, because they effectively forestall
of institutions and organizations to prevent the precise calculations of risk. Such barriers, by
harassment of individual environmental scientists delaying or impeding investigations, may not
who either attempt to investigate or report only block research, but they permit the
findings on hazard and risk which go against persistence of situations hazardous to the health
powerful interests that could be damaged by such and safety of the populations themselves. We see
information. use of the term “sound science” to disparage
The issues at hand here are not quite the attempts to make do with limitations of
same as those having to do with investigations of estimates of risk based on studies not meeting
scientific misconduct, i.e., the falsification or “gold standard” requirements because data sets
fabrication of research results. In investigations may not be complete. (38).
of scientific misconduct, there is a more or less A second form of harassment is lawsuits for
level playing-field for right and wrong: the peer libel. To address this hazard to environmental
reviewed literature and its well elaborated codes scientists, there is a need to explore the use of
and norms for evaluating scientific evidence. In insurance policies modeled after those available
the case of whistleblowing in environmental and to writers. Grants to environmental scientists
occupational epidemiology, the problem is to should include budgeting for such insurance.
promote access to this level playing field, and to
ensure that the playing-field is indeed level. Conclusions
There is a need to ensure that outside interests, Until now, there has been no watchdog address
often commercial, economic or political, do not
172
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science
for environmental and occupational an Epidemiological Study Used by EPA to Assess
epidemiologists to which to turn for assistance. Human Health Effects from Dioxins”, EPA
We suggest that major granting agencies memorandum to Raymond Loehr, Chairman, Executive
follow the lead of ISEE and the Collegium Committee, EPA Science Advisory Board, February 23,
1990. As cited in Rachel’s Hazardous Waste News #171.
Ramazzini in protecting environmental scientists Dioxins and Cancer–Fraudulent Stories. March 7, 1990.
from harassment. We call for studies on the 11. Jenkins C. (September 3, 1991) Affidavit: Recent
impact of harassment of research scientists on the Scientific Evidence Developed after 1984 Supporting a
detection and prevention of health risk. We call Causal Relationship between Dioxin and Human Health
for the development and use of codes for Effects. Full text of the affidavit is available online from
protecting environmental scientists from Agent Orange Victims, Families, and Friends in Touch.
harassment when they are engaged in this http://softvision.netfirms.com/ao_vets/.
mission. We recommend that measures to protect 12. Memo from Gerald H. Yamada, Office of [EPA’s]
messengers in environmental and occupational General Counsel, to Catherine L. Jenkins, dated
February 26, 1993. As cited in Rachel’s Hazardous
epidemiology be required of recipients of Waste News #328. An update on two key incinerator
research grants or contracts around the world. battles. March 11, 1993.
These codes should become a worldwide 13. Jenkins v. U.S. Environmental Protection Agency, 92-
institutional norm. Codes that protect CAA-6 (ALJ Dec. 14, 1992). Available online, http://
epidemiologic messengers in environmental and www.oalj.dol.gov/public/wblower/decsn/92caa06a.htm.
occupational medicine will serve also to protect 14. http://www.pmac.net/pesticide_wars.html. Pesticide
the public health. wars: the troubling story of Omar Shafey.
15. http://www.web.amnesty.org/ai.nsf/Index/
AMR250281997?OpenDocument&of=COUNTRIES%5CCUBA.
Acknowledgements Dr Desi MENDOZA Rivero, Cuba.
Councillors of ISEE and Collegium Ramazzini, 16. http://www.web.amnesty.org/ai.nsf/index/
which provided a grant for EDR’s work, and an EUR460421996 (1996) Federal Security Service (FSB)
anonymous donor for TB’s work. versus Prisoner of Conscience Aleksandr Nikitin: Return
to Soviet Practices.
Bibliography 17. www.rsf.fr/uk/html/europe/cplp01/lp01/261201.html -
Reporters wtihout Borders, Dec 2001 http://
www.internews.ru/ngo/pasko/indexe.html (1997)
1. Martin, B (1981) The scientific straightjacket: the power
Grigory Pasko sentenced to four years in prison.
structure of science and the suppression of
18. Martin B (1993) Stamping Out Dissent. Newsweek. 26
environmental scholarship. The Ecologist. Vol 11(1):
April: 49-50.
33-43.
19. Pesticide & Toxic Chemical News (1981) Dr. Mel
2. Deyo, RA, Psaty BM, Simon G, Wagner EH, Omenn GS
Reuber, Pathologist, Gets Sharp Censure, Warning from
(1997) The messenger under attack- intimidation of
his Supervisor (25 April): 22-23.
researchers by special-interest groups. New England
20. Watterson A (1998). Occupational health and safety
Journal of Medicine. 366 (16): 1176- 1180.
policy in the UK: who is driving the agenda with the
3. Relyea, HC (1994) Silencing Science: National Security
“greylisting” of health and safety activists? Int J Occup
Controls and Scientific Communication. Norwood, New
Environ Health 4(4):275-77.
Jersey: Ablex Publishing.
21. Coulter, J (1980) Memorandum re use of ethylene oxide
4. Hess, DJ (1999) Suppression, bias, and selection in
as a sterilant in the S.P.F. laboratory at Northfield, 16
science: the case of cancer research. Accountability in
April.
Research. 6: 245-257.
22. van den Bosch, R. (1978) The Pesticide Conspiracy.
5. Plato (1963) The trial of and death of Socrates. New
Garden City, NY: Doubleday.
York: Heritage Press.
23. http://home.cwru.edu/activism/kern.html. Conflicts of
6. Dror R. Notes on the trial of Socrates. Based on
interest: the case of David Kern.
conversations with the late Professor Yeshayahu
24. Westin, JB and Richter, ED (1990) The Israeli breast-
Liebowitz (personal correspondence, June 1999).
cancer anomaly. Ann NY Acad Sci 609:269-279.
7. Bentham, J (1907) An introduction to the principles of
25. Byers, RK and Lord, EE (1943) Late Effects of Lead
morals and legislation. Oxford: Clarendon Press.
Poisoning on Mental Development, American Journal of
8. Martin, B (1999) Suppressing research data: methods,
Diseases of Children. Vol. 6: 471-494.
context, accountability, and responses. Accountability in
26. Smythe, L (1980) letter, New Scientist, 85: 775-776. As
Research, Vol. 6: 333-372.
cited in reference Martin, B (reference 1).
9. Bahr, MH. “Violations of academic freedom: official
27. http://www.whistleblower.org/www/998news.htm#Oil.
statistics and personal reports”, Social problems, 14,
Big oil and the battle to protect the environment.
1967, pp. 310-320.
Government Accountability Project, Fall 1998.
10. Cate Jenkins, “Newly Revealed Fraud by Monsanto in
28. Sternberg, S (1996) An Unusual Case: Law Overseas
173
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Myron Mehlman’s case against Mobil Corp. attempts to
apply New Jersey statute to incident in Japan. The
Scientist 10(1).
29. De Blas, A. (1994) Environmental Effects of Mount
Lyell Operations on Macquarie Harbour and Strahan.
Sydney: Australian Centre for Independent Journalism,
University of Technology, Sydney.
30. http://www.jpost.com/Editions/2000/03/20/News/
News.4270.html. Siegel-Itzkovich, J (2000).
Environmental group aiding expert in libel suit.
Jerusalem Post Online Edition. March 20, 2000
31. http://www.iseepi.org/ethguide.htm#victimized. ISEE
support for victimized colleagues.
32. Abraham, J (1995). Science, Politics and the
Pharmaceutical Industry: Controversy and Bias in Drug
Regulation. London: ICL Press.
33. Alibek, KW and Handelman S (2000). Biohazard. Dell
Publishing Company.
34. Nigg, HN, Radulesc G (1994). Scientific Misconduct in
Environmental Science and Toxicology. JAMA
272:168-170.
35. LaFollette MC. Stealing Into Print: Fraud, Plagiarism,
and Misconduct in Scientific Publishing. Berkeley:
University of California Press; 1992:294.
36. American College of Epidemiology (2000). Ethics
Guidelines. Ann Epidemiol. 10: 487-497.
37. http://ori.dhhs.gov/html/policies/oripolicies.asp. Office
of Research Integrity.
38. Ong, EK and Glantz, SA (2001). Constructing “sound
science” and “good epidemiology”: tobacco, lawyers,
and public relations firms. Am J Public Health
91(11):1749-57

174
––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Richter, et al., Whistleblowers in Environmental Science

ISEE Ethics Committee Epidemiologist Whistleblower/Messenger Questionnaire:

1. Personal status
• ISEE Member? Y/N _____; ISEA Member? Y/N_____; Age _____
• Gender M/F _____
• Personal Status: M, S, D, W _____
• Children (Give no___ )

Education From To Where* (see Code) Code:

Undergrad ____ ____ __________ America: NA, CA, LA
MD ____ ____ __________ Euro pe: WestE, Med,
MPH/MSc/MS/MA ____ ____ __________ EastE Mideast: ME
PhD/DPH ____ ____ __________ Africa: WA, EA, SA Asia:
Post Doc ____ ____ __________ Ind, CentAsia, Jp, Ch,
Residency Spec ____ ____ __________ SEA Oceania: Aus, PI

2. Currently Employed
Where? _________________________ see code above
By: Govt Civilian Military Police (Circle one)
Level: National Regional/Province/District/Municipal (Circle one)
University/College ________________________
Independent research institute
Foundation _____
Trade Union NGO Self Employed
Industry/Corporation: If yes? ____________________
Multinational Y/N _____
Other_______

3. Tenured or permanent? Y/N

Rank (Univ): Prof Sr Lect/Lect__ Asst__Other_____

4. Research/salary funded by: (Circle correct answer)

Government
Industry
Foundation Other
No funding

5. Harassment: Following publication of research findings on health risks from environmental exposures,
have you ever experienced:
Ostracism Y/N Demotion Y/N Criminal investigation
/Prosecution/Trial Y/N
Confiscation of data Y/N Loss of job Y/N Physical threats Y/N
Threat of loss of job Y/N Threats of lawsuits Y/N Physical Attack Y/N
Transfer Y/N Lawsuits Y/N Imprisonment Y/N
Other

175
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

How many other co-researchers were there? ____ Did they experience any of the responses? Y/N

6. Research on specific problem which lead to episode(s) of harassment or threat or abuse: Years
during which research carried out: From______To_______________

Was this research on the hazard/risk published in:

Peer reviewed journal (sited in SCI CIT INDEX)___
Masters thesis___
Doctorate___
Internal document of organization in which you were then employed/studied?
Professional society___
Non peer-reviewed journal____
Other____
Date of publication?_______ Would you be able to provide the Citation? Leave blank if you wish
___________________________________________

7. Response
7a. Did you receive assistance after being subject to any of the above problems? Yes___
No____
7b. If yes, from: Individual colleagues____ _ Superiors____ Professional societies ___ NGO's
inside country_____ Journalists/Media_____ Lawyers or legal aid g roups____Colleagues
outside country____ NGO’s outside country_______Family______Other__________

8. Publication If findings were not published, were you prevented from submitting findings on health
risks on a hazardous exposure/risk for publication in a peer reviewed journal? Yes___ No____
OPTIONAL____________________________________________________________________

9. F indings: Could you summarize the findings you discovered/reported for which you were harassed?
Study design
(Cohort, CC, Prev, Pop(s) / N Exposure(s) Outcome RR/OR Reference
TS, Other)
_______________ __________ __________ __________ __________ __________
_______________ __________ __________ __________ __________ __________

10. In retrospect, were your findings: understated?____a proper assessment?____overstated?_____ For

further information: http//:www.iseepi.org/ethguide.htm

176
 II. Teaching

6. Training in the Responsible Conduct of Research

Influencing the Moral Dimensions of Professional Practice:
Implications for Teaching and Assessing for Research Integrity
Muriel J. Bebeau, Center for the Study of Ethical Development, University of Minnesota,
USA

Key Words: Assessment, Moral development, Professional ethics, Research ethics

This paper will present implications for teaching and assessing for research integrity from 20 years of
experience designing and assessing ethical development in the dental profession. Data sources for the
implications include: 1) pretest/posttest data for 18 cohorts of dental students who completed a well-
validated ethics program; 2) pre/post assessments of 28 practitioners referred by a licensing Board1
for individualized ethics instruction because they violated the State Dental Practice Act; and 3) efforts
in several professions to influence moral judgment development.
After pointing out some of the features of the Minnesota ethics program, the program’s
theoretical foundations (e.g., the processes of morality) are described. Each process suggests research
questions that motivate inquiry and assessment methods that were developed or used to investigate
the research questions and to gather evidence on program effectiveness. The paper continues with a
summary of data supporting the independence of the component processes and a discussion of the
ongoing search for behavioral indicators that could provide the “acid test” for the model. The paper
concludes with a discussion of the implications for the teaching and assessing for research integrity.
Special features2 of the curriculum include: 1) 43 contact hours distributed over four years;
2) required attendance and participation; 3) small group instruction—using dilemma discussion and
role-play; 4) an emphasis on student performance, self-assessment and personalized feedback; 5) use
of validated assessment methods that are checked for reliability; 6) involvement of high status
professionals (in measurement development and feedback); and 7) involvement of faculty in the
teaching. Thus, the curriculum isn’t a one-shot intervention, nor is it the isolated property of one
instructor.

Theoretical Foundations
The ethics curriculum, for students and referred practitioners, is designed to promote functional
processes that give rise to morality: 1) ethical sensitivity; 2) moral reasoning; 3) moral motivation
and commitment; and 4) ethical implementation (1). Moral failing is conceptualized as the result of
deficiencies in one or more of the processes. Rest’s Four Component Model of Morality,
operationally defined below, is a substantial departure from much of the work in psychology that
arbitrarily divides moral functioning into affects, cognitions, and behaviors (2).

The Four Component Model of Morality

Early in the cognitive developmental research program initiated by Kohlberg, he noted that, in
Corresponding author: Muriel J. Bebeau, Department of Preventive Sciences, School of Dentistry, University of Minnesota,
515 Delaware S.E., Minneapolis, MN 55455, 612-625-4633 (voice), 612-626-6096 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
addition to moral judgments, other processes components. Further, the Four Component
were important to the production of moral Model assumes that cognition and affect co-
behavior (3). Rest made these processes more occur in all areas of moral functioning. Thus,
explicit in what he called the Four Component moral action is not simply the result of separate
Model of Morality (1). Starting from the affective and cognitive processes operating in
question: how does moral behavior come about, interaction, as suggested by traditional models of
Rest suggested that the literature supports at least moral function that focus on three domains—
four component processes, all of which must be cognitions, affects and behavior (4, 5). Instead,
activated in order for moral behavior to occur. each of the four components are mixes of
These include: affective and cognitive processes that contribute
1. Moral sensitivity (interpreting the to the component’s primary function (e.g.,
situation as moral) identifying a situation as moral). Bebeau, Rest,
This process highlights the idea that moral & Narvaez suggest that researchers focus
behavior can only occur if the individual codes attention on identifying processes as they
the situation as moral. Specifically, Component contribute to moral action, rather than attempting
1 focuses on the various actions that are available to understand moral actions from a starting point
and how each action might affect the self and defined by arbitrarily dividing moral functioning
others. into affect, cognitions, and behavior (2).
2. Moral judgment (judging which of the The debate on the usefulness of a
available actions are most justified) psychological theory of morality, that has its
This is the process that Kohlberg emphasized. foundation in the work of Lawrence Kohlberg, is
Here the focus is on judging which of the various addressed in “Postconventional Moral Thinking”
options are the most ethically justified. Further, (6). This paper presents a theory of moral
the job of a psychologist and educator is in judgment development that is not grounded in a
sketching out how the justification process particularistic moral theory—as was
develops and under what conditions these Kohlberg’s—but is grounded in empirical
processes inform real-world choices. evidence illustrating that as individuals develop,
3. Moral motivation (prioritizing the so do the basic understandings they bring to
moral over other significant concerns) resolving complex moral problems. Such
Less understood than the other processes, the findings are of importance to ethics education in
main concern of Component 3 is, “why be general, as the goal of ethics education is, simply
moral.” The model acknowledges that put, to promote ethical development. The
individuals have a number of legitimate concerns authors contend that their findings will be of
that may not be compatible with the moral particular importance to research ethics educators
choice: for instance, career pressures, established because of their interest in promoting critical
relationships, idiosyncratic personal concerns, thinking about responsible research conduct (6).
among many others. Some of the most notable In the past, ethicists working in the professions
lapses of ethical behavior in the professions can questioned the usefulness of a moral
be attributed to low priority placed on the moral, development theory (and related measures) that
even when the moral choice is very well favored a particular moral theory, observing that
understood. practitioners working on real problems often
4. Moral character (being able to developed well-reasoned solutions without
construct and implement actions that service regard to a particular theory or even to
the moral choice) principlism as a way of arriving at moral
Component 4 represents the processes by which judgments (7).
one constructs an appropriate course of action, By amending a theory of moral judgment
avoids distractions, and maintains the courage to development to make it congruent with advances
continue. in moral philosophy, the authors hope to counter
It is important to notice that the model is not current views of the obsolescence of moral
conceived as a linear problem-solving model. psychology and support more interdisciplinary
For example, moral motivation may impact collaboration in the design and evaluation of
moral sensitivity, and moral character may moral education programs. Further, a more
constrain moral motivation. In fact, Rest (1) enlightened view of the role of tests of moral
makes clear the interactive nature of the judgment development should enable educators
180
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
to put such tests to more appropriate use. description of each measure and the findings are
Besides drawing attention to a broader summarized as follows:
conception of postconventional moral thinking,
the authors direct the reader’s attention to a Component I: Ethical Sensitivity
broader conception of morality, one that
encompasses moral judgment, but that also The Dental Ethical Sensitivity Test (DEST)
addresses other aspects of moral functioning, The DEST (Form A or B) (10, 11) assesses the
including moral sensitivity, motivation, character, ability to recognize the ethical issues hidden
and competence. The Four Component Model of within the professional problems dentists
Morality has been a centerpiece for research encounter in practice. Students’ verbal responses
activities at the Center for the Study of Ethical to four audio-taped dramas are recorded and
Development for nearly 20 years. transcribed, and provided to the student and to a
practicing dentist, who each apply the DEST
coding scheme, then meet for personalized
Educational Interventions Assessed in
feedback. The validity and reliability of the
Terms of the Four Components DEST are reported in several studies,
A program of research and educational summarized in Bebeau (8) and Fravel and
development to investigate the usefulness of the Bebeau (12). Briefly, the results support these
model was initiated by Jim Rest and the author in conclusions: 1) Ethical sensitivity can be
the early 80s. Variations on these research reliably assessed. Calibrated raters achieved item
questions motivated the inquiry: Can ethical agreement ranging from 84.7 percent to 88
sensitivity (or any of the other components) be percent. Reliability estimates for individual
reliably assessed? Do students differ in ethical cases ranged from .83 to .92; 2) Students and
sensitivity (or other components)? Can practitioners vary in sensitivity to ethical issues.
sensitivity (or other components) be enhanced? Students at different levels of education in
And, is ethical sensitivity distinct from other medicine and dentistry (physicians vs.
components? technicians or dentists vs. hygienists) differed
The Four Component Model offers unique significantly, such that those with longer
information and direction for educational preparation showed higher levels of sensitivity.
development. First, it suggests profitable areas Further, the DEST is sensitive to institutional
for measurement development. To claim that a differences; 3) Women have a slight edge over
program is effective in a broad sense, it seems men in recognizing ethical issues, but differences
reasonable to expect changes within each of the were not attributed to differential recognition of
four components. For the dental curriculum, the care and justice issues; 4) Ethical sensitivity
measures of each component were designed and can be enhanced through instruction; 5) Ethical
validated, and data from them helped identify sensitivity is distinct from moral reasoning
deficiencies to consider as the curriculum was abilities. Correlations between the DEST and
designed. There are measurement models and Defining Issues Test (DIT) posttest are
methods for assessing each of the components (2, consistently low (see later section for more
8). These can be used as templates for detail); 6) Despite the stressful nature of the
assessment in various contexts. DEST assessment—responding on the spot to
Second, the model provided direction for complex cases, having responses taped,
instructional design for groups, as well as for transcribed, and sent to a practicing professional
individual referrals. For referred practitioners, is a high-stakes examination—students value the
deficiencies were noted in various components assessment and feedback experience.
and were associated with particular moral
weaknesses (9). Targeting specific deficiencies
Component II: Moral Reasoning and
in an individualized instructional program proved
to be an effective intervention strategy, resulting Judgment
in substantially enhanced posttest performance. In this section, two measures are described: a
well-established measure (DIT) and a newly-
Measures for the Components of Morality devised, context-specific test of ethical reasoning
Five measures are used to assess performance in and judgment (Dental Ethical Reasoning and
the Dental Ethics Curriculum. A brief Judgment Test [DERJT]). In the case of the DIT,

181
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
the discussion will include findings from new classes tested). The average graduate selects
analyses with new indices for three of the recent postconventional arguments 51 percent of the
cohorts of dental students. time (with cohorts ranging from 47 to 55). Effect
The Defining Issues Test sizes vary across classes, with a range of .12 to
The DIT measures life-span development of .78, with an average of .43. For each cohort,
moral reasoning and judgment (13). The DIT is scores tend to be normally distributed. For
the most widely used test of moral judgment entering students, as many as 35 percent are not
development and is often used as an outcome using postconventional moral schemas as often
measure for intervention studies, because it has as the average adult, with about seven percent
an exceptional validation history.3 Students read above the mean of philosophy and political
dilemmas, and then rate and rank the importance science graduate students. Although we see an
of each of 12 arguments to support their position. upward shift in the distribution at posttest, with
Confirmatory factor analysis of a mega-sample 16 percent lower than the mean of the average
of over 44,000 subjects shows that items adult, and 20 percent above the mean of
(arguments) cluster around three general moral philosophy and political science graduates; of
schemas: Personal Interest, Maintaining Norms, particular interest are the proportion of students
and Postconventional schemas (14). Typically, who showed no change or regressed from pretest
researchers have reported scores in terms of the P to posttest. By classifying students’ change
score—the proportion of items selected that scores into categories defined by the standard
appeal to Postconventional moral frameworks for error of measurement (18), Bebeau reported that
making decisions. The average adult selects 44 percent of the 1,229 students who participated
postconventional moral arguments about 40 in the curriculum made moderate to highly
percent of the time, the average Ph.D. candidate significant gains, 40 percent showed no change,
in moral philosophy or political science about and 16 percent regressed on the P score (20).
65.2 percent of the time, the average graduate New Indices and New Analyses of DIT
student 53.5, with the average college graduate at Scores
42, and the average high school student at 31.8 Observations of what appeared to be regression
percent. in postconventional reasoning in our intervention
Progress in moral judgment is studies prompted the validation studies, including
developmental, and development proceeds as development of an alternate form of the DIT and
long as an individual is in an environment that a reanalysis of moral education interventions that
stimulates moral thinking. College has a attended to several moral cognition variables
powerful effect on moral judgment development. derived from DIT scores (6, 14, 21, 22, 23, 24).
McNeel’s meta analysis of 22 longitudinal Moral Schema Profiles. Instead of relying
studies of liberal arts students estimates first year only on the P score as a measure of pretest to
college students at 36, seniors at 46, estimating posttest change, a profile showing the proportion
an effect size of .80 (15). Effect sizes of about of times a student rates was constructed to
0.80 are among the largest effect sizes for many illustrate important items for each of three
college impact variables that have been studied. general schema: a Personal Interests schema
In fact, effect sizes are higher for moral judgment (Kohlbergian Stage 2 and 3 items); a Maintaining
than for the many cognitive and affective college Norms schema (Stage 4 items): and a
outcome variables that have been studied (16). Postconventional schema (Stage 5 and 6 items).
Yet professional schools (e.g., Veterinary Figure 1 illustrates how two profiles with similar
Medicine, Medicine, Dentistry, and Accounting) P scores can reflect differing levels of moral
are programs where one does not typically see judgment development. Examining profiles from
gains associated with the educational program, students who did not show gains in DIT P scores
unless the program has a specially-designed from pretest to posttest (20) illustrates a
ethics curriculum (17). Further, for some substantial reduction on the Personal Interest
students and some professions, programs actually schema coupled with an increase on the
seem to inhibit growth (18, 19). Maintaining Norms schema, without significant
Change in moral judgment can be attributed change on the Postconventional schema score. In
to the ethics curriculum (18). The average fact, when the statistically significant pretest/
entering Minnesota dental student scores 46 posttest change for the 18 cohorts of students that
(with cohorts ranging from 42 to 49 across the 15 participated in the dental curriculum was
182
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice

50
peaking at Type 6. For the posttest responses,
PI MN P 75.8 percent were classified at Types 5 and 6,
45
Type 4 with 59.9 percent at Type 6. By way of
40
comparison, Yeap reported college student
35 samples peaked at Type 3.
30 These new analytical procedures may help to
25
unravel some of the puzzles researchers have
cited, where professional groups like Accounting
20
and Auditing (19) seem to regress on moral
15
judgment as a result of an educational program.
10 Such analysis may clarify McNeel’s findings that
5 programs that are too careerist (focus narrowly
0
on technicalities of beginning job performance)
Transition
Transition
Type 3
Type 3
Consolidation
Consolidation
Type 4
or too dogmatic (in closing off questioning and
inquiry) inhibit growth in reasoning (15). Such
Figure 1. Moral judgment profiles illustrating similar P
findings would have implications for developing
scores, but differences on other moral cognition variables.
PI = Personal Interests Schema
research integrity. Courses that focus narrowly
MN = Maintaining Norms Schema on the rules of research conduct may focus
P = Postconventional Moral Schema attention on the minimal (legal) standards, rather
than on aspirational standards for research
reanalyzed, the reduction in the Personal Interests integrity.
schema score appeared much greater and more Tests like the DIT are valuable for assessing
consistent across cohorts than changes in P score. general reasoning that is a critical element of
By focusing only on the P score, researchers may professional ethical development, but they may
be missing change that is quite significant. not be sensitive to the specific concepts taught in
Consolidation/Transition. Figure 1 a professional ethics course—or indeed, in a
illustrates another variable to consider in research ethics course. The question (for
describing change. When there is little evidence educators) is often whether to teach specifically
of discrimination among the schema-typed items, to the codes or policy manuals, or to teach
students are classified as transitional. A flat concepts particular to a discipline—informed
profile is viewed as a marker of developmental consent, intellectual property, conflict of interest,
disequilibrium, or transition, since there is no etc.
evidence of a schema preference. A further The Dental Ethical Reasoning and
discussion of this topic is addressed by Thoma Judgment Test (DERJT)
and Rest (22). A pretest/posttest analysis of The DERJT is a first effort to test application of
consolidation/transition status was conducted for context-specific concepts (taught in ethics
222 dental students (20), showing that nearly half courses) to real cases (25). The test is similar to
the students (46.9%) were in a transitional status the DIT, in that cases are presented followed by
at pretest, whereas only 27.1 percent exhibited lists of action choices and justifications. The
the transitional status at posttest. action and justification choices for each problem
Type. Profiles can further be classified by were generated by a group of Minnesota dental
type (22), where type reflects both the faculty and residents. The scoring key was
predominant schema and the extent of its use. developed by a group of “dental ethical experts.”
By reexamining several intervention studies When taking the test, a respondent rates each
reported in the literature, Yeap showed that Type action or justification, then selects the two best
provided a more illuminating description of and two worst action choices, and the three best
change that occurred as a result of an and two worst justifications. Scores are
intervention than relying simply on the P Score determined by calculating the proportion of times
(24). A pretest/posttest analysis of six Types was a respondent selects action choices and
also conducted for the 222 students reported justifications consistent with “expert judgment.”
above. Whereas the pretest responses were In validation studies, Bebeau and Thoma have
distributed among Types 3, 4, 5, and 6, 61.2 seen clear expert novice differences (25).
percent were classified at Types 5 and 6 Further, scores for students, practitioners, and
(postconventional types), with the distribution referrals appear to be normally distributed. In a
183
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
study comparing our graduates’ responses to of professionalism are included to stimulate
familiar vs. unfamiliar problems presented on the thinking about the model of professionalism that
test, it appears that a good grasp of appears to be dominant for the individual. When
postconventional moral schemas is a necessary the scales and interpretive guide are used in an
condition for transfer to new problems. educational setting, participants can compare and
discuss items and challenge each other’s
Component III: Motivation and thinking.
Commitment Developing a concept of role appears to
require instruction and opportunities for
The Professional Role Orientation reflection. At entry to professional school,
Inventory (PROI) Minnesota dental students do not illustrate a
The PROI assesses commitment to privilege good understanding of key concepts of
professional values over personal values (26, 27). professionalism like service to society, or the
Likert scales assess dimensions of priority of patient well-being, or the duty to self-
professionalism that are theoretically linked to regulation (8). But, even after participation in an
models of professionalism described in the instructional program in which students write an
professional ethics literature. The PROI scales, essay describing their perception of their
in particular the responsibility and authority professional role (the program is of demonstrated
scales, have been shown to consistently effectiveness and includes generous amounts of
differentiate beginning and advanced student practice and feedback on performance), key
groups and practitioner groups expected to differ concepts like self-regulation, service to society,
in role concept. By plotting responses of a and the basic duty to place patient’s rights before
cohort group on a two dimensional grid, four self-interest are still frequently omitted or
distinctly different views of professionalism are miscommunicated by as many as 20 percent of
observed (26) and, if applied, would favor the students. The literature on concept learning
different decisions about the extent of has helped us see that when students have no
responsibility to others. In comparing practicing functional schema for a particular concept,
dentists with entering students and graduates, our several educational experiences are required to
graduates consistently express a significantly instill a clear concept of the professional’s role.
greater sense of responsibility to others than Whether instilling a clear idea of the
entering students and practicing dentists from the professional’s role will motivate students to place
region. This finding has been replicated for five moral values over personal ones is a key
cohorts of graduates (n = 379). Additionally, the question. The most direct evidence of a
graduates’ mean score was not significantly relationship between role concept and
different from a group of 48 dentists, who professionalism comes from the study of
demonstrated special commitment to performance of the 28 members of the practicing
professionalism by volunteering to participate in community, referred for courses in dental ethics
a national seminar to train ethics seminar leaders. because of violations of the dental practice act.
A recent comparison of pretest/posttest scores for Although the practitioners varied considerably on
the Classes of 1997-1999 (20) indicates measures of ethical sensitivity, reasoning, and
significant change (p < .0001) from pretest to ethical implementation, 27 of 28 were unable to
posttest. Cross-sectional studies of differences clearly articulate role expectations for a
between pre and posttest scores for a comparable professional (9).
dental program suggests that ethics instruction
accounts for change. Component IV: Moral Implementation
To provide students or practitioners with
(character and competence)
individualized feedback on their role concept, an
interpretive guide is provided enabling a Shifting to the last component, character and
respondent to sum his or her own scores on each competence, the authors have observed that
scale, plot them on the two dimensional grid (one guided practice changes the expectation of
grid is provided for the authority and efficacy that is likely to change behavior. Role-
responsibility scales, one for the agency and playing builds competence and confidence in
autonomy scales), and then compare responses to resolving thorny ethical problems, and skills in
their cohort. Descriptions of each of the models

184
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
communication and negotiation are necessary moral judgment is linked to a wide range of pro-
requisites of this competence. social behaviors (28), including clinical
A Professional Problem Solving Index performance ratings for nurses (29, 30),
Problem-solving and role-playing performance physicians (31) and dentists (8), and to
scores are calculated for eight complex cases that preferences for the more altruistic law disciplines
present difficult human interaction problems (8, for law students (32), the search for behavioral
20). Students are directed to prepare 1) an measures to examine the relative contribution of
interpretation of the facts that must be addressed each component to the behavioral outcomes has
if the problem is to be resolved efficiently; 2) an been a frustrating one. The author’s most recent
action plan; and 3) a verbatim dialog to illustrate effort (20) has been to calculate a productivity
the implementation of the action plan. A index that reflects students’ success in interacting
checklist, prepared for each case, assures some effectively with patients to achieve acceptance
uniformity in judging responses. Each response and completion of treatment recommendations.
is reviewed by a peer and by the course instructor To meet competency requirements, the student
who provide written comments identifying the must achieve an average monthly index (over all
strengths and shortcomings of the assignment. months of clinical practice) of .75 or above.
As with other measures, scores are normally Although there was considerable range in
distributed and cohort differences are observed. productivity from .67 to 1.19, since students must
meet a .75 overall average in order to graduate,
Independence of the Components of the productivity index, while identifying highly
Morality effective students, also produces a highly skewed
distribution (Mean = .80, S.D. = .08). In the
Rest’s Four Component Model predicts the analysis, productivity, like Grade Point Average,
independence of the components (1). Prior was not related to any of the measures of
studies have typically reported low to very low morality.
correlations between ethical sensitivity and moral The explanatory power of the Four
judgment, but correlations among the other Component Model is observed, taking a
components have varied from very low to an somewhat different approach, i.e., working
occasional moderate correlation. Often sample backward from disciplinary action to examining
sizes have been low, challenging the reliability of deficiencies in the components. Baldwin
the estimates. Recently, Bebeau reported observed a relationship between the number of
correlations between components for a larger malpractice claims and moral judgment scores,
sample (230 students) (20). Except for the noting that a high DIT score had a kind of
expected moderate correlations (.46) between the protective effect, insulating one from claims (33).
DIT Pretest and Posttest and between the PROI For dental practitioners referred for ethics
Pretest and Posttest scales (.38), each measure instruction, disciplinary actions were directly tied
appears to provide unique information about to significant deficits in one or more of the
ethical decision making competence. Consistent components (8, 9). Further, one consistent
with earlier studies, correlations are consistently observation, in addition to a deficiency in either
very low between the DEST and the DIT, and sensitivity, reasoning or implementation, is the
between the DEST and other component difficulty 27 of the 28 referrals had in articulating
measures (8). The exception is between the the expectations of the profession. After targeted
DEST and the DERJT justification score, where instruction, directed toward role concept
there appears to be some overlap between the development and remediation of one or more
two tests (r = .28). Also consistent with earlier other deficiencies, we observed measurable
reports (27), there appears to be some low to improvements in performance, coupled with
moderately-low relationship between the PROI documented changes in the behaviors that gave
Responsibility Scales and the DEST and DIT. rise to the disciplinary action. Further, to date,
there have been no cases of recidivism.4
The Continuing Search for Behavioral Examining case studies bolsters the
Indicators understanding of the connection between the
Several attempts have been made to show the components and behavior, and provides direction
contributions of each of the components to for education.
meaningful behavioral indicators. Although
185
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Conclusions sometimes much more difficult.

Analyzing data from the sources cited indicates: The author’s experience suggests that for
1) striking individual differences among students novice ethics teachers (which most of us are)
and practicing professionals on each of the focusing on sensitivity, reasoning, and role
measures; 2) that competence on one of the concept independently of one another will more
processes does not predict competence on efficiently develop the skill needed for effective
another; 3) that curricula of rather modest problem solving. Ethics teachers should not
duration can influence performance in expect that carefully targeted courses will
measurable ways (our curriculum consists of 43 develop the more advanced skills in ethical
contact hours); and 4) that strengths and reasoning that might result from courses in moral
weaknesses in each of the processes are linked to philosophy. Yet, problem-based practice (using
real-life ethical behavior. The findings described cases) can be especially effective in helping
in this paper support Rest’s contention that moral students recognize and subsequently avoid
failings can result from deficiencies in one or personal interest arguments while strengthening
more of the processes. Findings also support the awareness and adherence to the rules of
importance of attending to each when designing responsible research conduct.
curriculum. Further, whether a curriculum
promotes ethical development depends on Notes
whether that curriculum incorporates the 1. The referrals from the State Board came about because
elements of effective instruction. some of the Board members have been involved in the
undergraduate curriculum for students. They wondered
whether violations of the Dental Practice Act reflected
Implications for Teaching and Assessing ethical deficiencies that could be remediated by the
for Research Integrity kinds of experiences we provided for students.
If the objective is to develop thoughtful and 2. For a detailed account of the undergraduate dental ethics
responsible scientists who act with integrity and curriculum, see Bebeau (1994).
have broad understanding of their role and a 3. There is extensive literature on the construct validity of
commitment to integrity in science, it is the DIT. See Rest, Narvaez, Bebeau, & Thoma (1999)
important to do more than teach the rules and for a summary and references to the 400 published
policies that apply to the conduct of research. studies using the DIT.
4. It is important to note that none of the practitioners
Before engaging in case discussions, research
referred for remediation involved problems with
ethics teachers need to address the expectations impulse control, substance abuse, mental illness, or
of a scientist. Students cannot be expected to significant personality disorders.
intuit the norms and values that undergird the
research enterprise. And, it is not clear that they Bibliography
can “pick them up” from role models. The 1. Rest J. Morality. In: Mussen PH, series editor, Flavell J,
expectations need to be explicitly taught and Markman E, volume editors. Handbook of child
formally assessed, preferably in writing. By psychology: vol. 3, cognitive development. 4th ed. New
asking students to express the concepts in their York: Wiley; 1983. p. 556-629.
own words, and in writing, misperceptions can 2. Bebeau MJ, Rest JR, Narváez DF. Beyond the promise:
be identified and addressed before they become a perspective for research in moral education.
an issue. Once the expectations of the scientist Educational Researcher 1999;28 (4):18-26.
3. Kohlberg L. Stage and sequence: the cognitive
are clear, engage students in active learning developmental approach to socialization. In: Goslin DA,
(using cases, if possible) to facilitate ethical editor. Handbook of socialization theory. Chicago: Rand
sensitivity, reasoning and problem solving. McNally; 1969. p. 347-80.
When designing case materials, careful thought 4. Eisenberg N. Altruistic emotion, cognition and behavior.
should be given to the particular process that is Hillsdale (NJ): Erlbaum; 1986.
of concern. Too often, cases are written and 5. Lickona T. Educating for character. New York: Bantam;
participants are asked: What should the 1991.
protagonist do? Such a question focuses on 6. Rest J, Narvaez D, Bebeau MJ, Thoma SJ.
problem solving, rather than problem Postconventional moral thinking: a neo-Kohlbergian
approach. Mahwah (NJ): Erlbaum; 1999.
identification or moral reasoning. Certainly a 7. Beauchamp TL, Childress JF. Principles of biomedical
skilled facilitator can redirect attention to ethics. 4th ed. New York: Oxford University Press;
reasoning or problem identification, but it is 1994.

186
–––––––––––––––––––––––––––––––––––––––––––––––– Bebeau, Influencing the Moral Dimensions of Professional Practice
8. Bebeau MJ. Influencing the moral dimensions of dental 24. Yeap CH. An analysis of the effects of moral education
practice. In: Rest JR, Narvaez D, editors. Moral interventions on the development of moral cognition
development in the professions: psychology and applied [dissertation]. Minneapolis (MN): University of
ethics. Hillsdale (NJ): Erlbaum; 1994. p. 121-46. Minnesota; 1999.
9. Bebeau MJ. Continuing ethics education for 25. Bebeau MJ, Thoma SJ. “Intermediate” concepts and the
remediation. Paper presented at the meeting of the connection to moral education. Educational Psychology
American Association of Dental Schools; Orlando (FL); Review 1999;11 (4):343-60.
1997. 26. Bebeau MJ, Born DO, Ozar DT. The development of a
10. Bebeau MJ, Rest JR, Yamoor CM. Measuring the ethical Professional Role Orientation Inventory. Journal of the
sensitivity of dental students. J Dent Educ 1985;49 American College of Dentists 1993;60 (2):27-33.
(4):225-35. 27. Thoma SJ, Bebeau MJ, Born DO. Further analysis of the
11. Bebeau MJ, Rest JR. The Dental Ethical Sensitivity Professional Role Orientation Inventory [abstract 116]. J
Test. Minneapolis (MN): Center for the Study of Ethical Dent Res 1998 Mar;77 (Special Issue):120.
Development; 1990. 28. Thoma S. Moral judgments and moral action. In: Rest
12. Fravel BJ, Bebeau MJ. The role of the American JR, Narvaez D, editors. Moral development in the
College of Dentists in assessing ethical sensitivity. professions: psychology and applied ethics. Hillsdale
Poster session presented at the 25th annual meeting of (NJ): Erlbaum; 1994. p. 199-212.
the Association of Moral Education; Minneapolis (MN); 29. Duckett LJ, Ryden MB. Education for ethical nursing
1999 Nov. practice. In: Rest JR, Narvaez D, editors. Moral
13. Rest J. Development in judging moral issues. development in the professions: psychology and applied
Minneapolis (MN): University of Minnesota Press; ethics. Hillsdale (NJ): Erlbaum; 1994. p. 51-70.
1979. 30. Krichbaum K, Rowan M, Duckett L, Ryden MB, Savik
14. Rest J, Thoma S, Edwards L. Designing and validating a K. The clinical evaluation tool: a measure of the quality
measure of moral judgment: stage preference and stage of clinical performance of baccalaureate nursing
consistency approaches. Journal of Educational students. J Nurs Educ 1994;33 (9):395-404.
Psychology 1997;89 (1):5-28. 31. Sheehan TJ, Husted SD, Candee D, Cook CD, Bargen
15. McNeel SP. College teaching and student moral M. Moral judgment as a predictor of clinicial
development. In: Rest JR, Narvaez DF, editors. Moral performance. Evaluation and the Health Professions
development in the professions: psychology and applied 1980;3:393-404.
ethics. Hillsdale (NJ): Erlbaum; 1994. p. 27-50. 32. Landsman M, McNeel SP. Moral judgment and
16. Pascarella ET, Terenzini PT. Moral development. In: preference for public interest law practices among
How college affects students: findings and insights beginning law students. Paper presented at the Annual
from twenty years of research. San Francisco (CA): Meeting of the Association for Moral Education;
Jossey-Bass; 1991. p. 335-68. Glasgow, Scotland; 2000 July 7-11.
17. Rest JR, Narvaez D, editors. Moral development in the 33. Self DJ, Baldwin DC Jr. Moral reasoning in medicine.
professions: psychology and applied ethics. Hillsdale In: Rest JR, Narvaez D, editors. Moral development in
(NJ): Erlbaum; 1994. the professions: psychology and applied ethics.
18. Bebeau MJ, Thoma SJ. The impact of a dental ethics Hillsdale (NJ): Erlbaum; 1994. p. 147-62.
curriculum on moral reasoning. J Dent Educ 1994;58
(9):684-92.
19. Ponemon L, Gabhart D. Ethical reasoning research in
the accounting and auditing professions. In: Rest JR,
Narvaez D, editor. Moral development in the
professions: psychology and applied ethics. Hillsdale
(NJ): Erlbaum; 1994. p. 101-19.
20. Bebeau MJ. Influencing the moral dimensions of dental
practice: implications for the professions. Invited
address. American Educational Research Association;
New Orleans (LA); 2000 Apr 24-28.
21. Rest JR, Thoma SJ, Narvaez D, Bebeau MJ. Alchemy
and beyond: indexing the Defining Issues Test. Journal
of Educational Psychology 1997;89 (3), 498-507.
22. Thoma SJ, Rest JR. The relationship between moral
decision-making and patterns of consolidation and
transition in moral judgment development.
Developmental Psychology 1999;35:323-33.
23. Rest J, Narvaez D, Thoma SJ, Bebeau MJ. DIT2:
devising and testing a revised instrument of moral
judgment. Journal of Educational Psychology 1999;91
(4):644-59.
187
Research Ethics in US Medical Education: An Analysis of Ethics Course
Syllabi
James M. DuBois, Center for Health Care Ethics, Saint Louis University, USA
Jill E. Ciesla, Center for Health Care Ethics, Saint Louis University, USA
Kevin E. Voss, Center for Health Care Ethics, Saint Louis University, USA

Keywords: Curriculum, Ethics, Medical education, Medical ethics, Medical schools

Medical education trains future physicians as medical practitioners. For this reason ethics education
for medical students has traditionally focused on themes revolving around the patient-physician
relationship: veracity, informed consent, fidelity, confidentiality, non-maleficence, and the like (1-3).
While many of these themes overlap with themes in research ethics, these ethics courses may be
inadequate for those future physicians who will engage in research of any kind – including clinical
trials, patient surveys, or program assessments (4-7). Research ethics introduces new and important
themes related to experimental design, interaction with communities, and the dissemination of
information (8,9). The well being of patients, physicians, and research institutions is at stake when
physicians fail to abide by rules for ethical research (9,10).
Recent, highly publicized failures to follow protocol at major medical centers reinforce the idea
that Institutional Review Boards (IRBs) are inadequate to ensure ethical research behavior. These
facts give rise to an important research question: To what extent is research ethics incorporated into
the ethics curriculum at medical schools in the United States (US), where future clinical researchers
are trained? This question takes on additional significance when one considers that medical students
may be engaged in clinical research in various forms even before completing undergraduate medical
studies (5,11,12).
This study builds upon a larger study that the first two authors of this paper conducted on the
ethics curriculum in US medical schools. DuBois and Ciesla analyzed syllabi from required ethics
courses in US medical schools with the aim of identifying and rank-ordering course objectives,
teaching methods, course content, and methods of student assessment (13). (The term “ethics course”
is used here to refer broadly either to a self-standing course or to a formal educational unit within a
larger course.) The present study analyzes in detail the content of the research ethics portion of
required ethics courses in the 4-year medical doctor (MD) curriculum at US medical schools. It
makes no attempt to describe responsible conduct of research (RCR) education at medical schools as
a whole, which frequently house graduate and postgraduate programs in the biomedical sciences, and
accordingly offer more extensive RCR courses outside of their MD programs.

Methods
This study was presented to the Institutional Review Board of Saint Louis University. It was approved
Corresponding author: James M. DuBois, PhD, DSc, Center for Health Care Ethics, Saint Louis University, Salus Center,
4th Floor, St. Louis MO 63104, 314-977-6660 (voice), 314-268 5150 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
as an exempt study given guarantees that (n=87). Seventy-nine percent (n=69) of these
participation would be voluntary, subjects would schools claimed to require a formal ethics course.
be adults, and confidentiality would be Of these schools, 84% (n=58) provided ethics
maintained by publishing only aggregated data. course syllabi. The two raters categorized items
the same in 90% of the cases. In the predecessor
Instrument and Participants study, analysis and codification of all syllabi
The American Association of Medical Colleges identified 10 course objectives, 8 teaching
(AAMC) provided mailing labels for all methods, 39 content areas, and 6 methods of
curriculum directors of 4-year medical colleges student assessment. The mean for individual
in the US (N=121). A 1-page survey was sent to schools was 3 objectives, 4 teaching methods, 13
all curriculum directors asking whether ethics is content areas, and 2 methods of assessment.
taught as a formal required component, as an Among the 39 different content areas,
elective, or not at all. It also inquired into the research ethics ranked 11th. Twenty-three of the
year or years in which ethics is taught. The 58 syllabi (39.6%) addressed research ethics in
survey further requested course syllabi for all some fashion. Analysis of the research ethics
formal ethics components in the 4-year medical sections of these syllabi revealed 82 specific
curriculum. themes that fall under 17 different general
categories.
Analysis Table I (below) presents these 17 general
In the larger study, two researchers read all categories in rank order, along with the specific
syllabi using an open coding method to produce a themes that fall under each category. It further
comprehensive list of all elements found in the indicates where the categories and specific
syllabi that fell into one of four generic themes overlap with the US Public Health
categories: (1) course objectives, (2) teaching Service’s (PHS) “Core Instruction Areas” for
methods, (3) course content, and (4) student courses on the Responsible Conduct of Research
assessment methods. All other statements (e.g., (RCR) (14). (This policy of December 1, 2000
pertaining to class times, locations, and was suspended by the Bush administration in
instructors) were ignored. The specific elements February 2001 pending further study. This paper
of the syllabi were then placed into categories. refers to the policy because it continues to serve
These categories were used to create variables in as a model for many institutions and it remains
a SPSS database. Schools, rather than syllabi, under discussion among legislators and policy
constituted cases in the database: if a school had makers.)
more than one required ethics component, data The average number of general research
from all required course syllabi were entered into ethics topics addressed in these 23 syllabi is 6,
that case. Data from 10 syllabi (17%) were with individual schools covering anywhere from
entered by two researchers to establish interrater 1 to 11 topics. Only six topics were covered by
reliability. more than half of those syllabi that address
The present study identified those syllabi that research ethics. In rank order these are: clinical
included content on research ethics. trials; informed consent; general ethics of human
The research ethics sections of syllabi were subject research; government committees and
read using an open-coding method to generate a regulations; history and background to research
comprehensive list of research ethics content. ethics; and protecting vulnerable populations. No
The results of this open-coding process were then research ethics topic was covered by more than
placed into general categories. These categories 21% of the 87 participating schools. The number
were entered into an expanded SPSS database. of research ethics topics covered did not correlate
Statistical analysis aimed above all to provide significantly with either school enrollment
descriptive data on the frequency of various (r=.10, p<.45) or tuition costs (r=.10, p<.43).
research ethics content. Pearson’s r was used to
test whether the mean number of content areas Discussion
covered was significantly correlated with either While Mastroianni and Kahn conducted a useful
class size or tuition cost. and informative pilot study of NIH grantee
institutions’ training efforts in RCR, this study is
Results the first to examine comprehensively the RCR
Surveys were returned by 72% of the schools curriculum in US medical programs. Our study
190
–––––––––––––––––––––––––––––––––––––––––––––––––––––––– DuBois, et al., Research Ethics in US Medical Education
exposes two possible causes for concern. First, undergraduate medical students in research
too few medical schools teach research ethics in (5,11,12). However, the prevalence and extent of
any fashion within their MD program. No topic students’ roles and whether they are specifically
in research ethics – including clinical trials – is listed in study protocols remains unknown. Thus,
covered by more than 21% of all medical it is difficult to know whether education in RCR
schools. The topic of Institutional Review Boards is a pressing need for medical students, or
is covered by less than 13% of medical schools, whether these years might be viewed simply as a
despite the fact that medical researchers are most convenient time to introduce education in RCR.
likely to work precisely with human subjects. Research has shown that private industry is
Second, it appears that important topics are now funding more research than is the
wholly missing even in those programs that teach government (15). Government requirements
research ethics. This becomes clear when regarding RCR instruction pertain only to
comparing the specific research ethics topics government-funded research, and according to at
covered within medical ethics syllabi to the least one study, two-thirds of NIH grantee
“Core Instruction Areas” PHS identified for RCR institutions require RCR instruction only to the
education (14). For example, the first five of nine extent that the government mandates it (16).
core areas PHS identifies (data acquisition, These facts suggest that a “blanket” approach to
management, sharing, and ownership; mentor / educating future physicians would be the safest
trainee responsibilities; publication practices and route to ensuring RCR instruction for clinical
responsible authorship; peer review; and researchers. However, given the scope of recent
collaborative science) seem wholly missing from government requirements, such a blanket
these syllabi. (The only possible exception is one approach would have to be initiated by a
syllabus that mentions industry/university professional institution like the AAMC.
relationships.) Finally, it is difficult to anticipate how well
It is, of course, possible that some of these the RCR programs that are currently being
topics are covered under other general headings mandated will address the specific ethical
(e.g. ‘collaborative research’ might be discussed concerns that arise in clinical, medical research.
under ‘clinical trials’). This is one limitation of This study has shown that 13 of our 17 categories
the method used: a topic is identified only if it could easily be subsumed under just one PHS
explicitly appears on the course syllabus. This Core Area: #6, Human Subjects. This suggests
means that syllabi using only very general that typical RCR instruction aims to cover a
headings will be shortchanged. Nevertheless, a broad range of issues that arise in research (such
course syllabus should be a reliable statement of as authorship, peer review and the treatment of
the objectives and content of a course, and most animals), whereas physicians feel the need for a
syllabi were quite detailed (as the larger study highly focused and intensive treatment of human
demonstrated). Thus, it seems safe to conclude subject research. The years of medical school
both that very few MD programs discuss research may be the best or only time to provide this sort
ethics and that those that do ignore at least half of of special-tailored education in RCR.
the topics PHS wants to see addressed. While this study has provided new answers
However, the significance of these findings to questions about the current educational
cannot be firmly established until other questions training of medical students in RCR, it has also
are answered: managed to bring new questions to the fore. Only
• To what extent are medical students partici- after these questions are answered, will the
pating in clinical research? significance of this study’s findings be properly
• Are current requirements for RCR instruction understood.
likely to be successful in targeting future
physicians who are funded by private Acknowledgements
The authors thank Doris Rubio, Ph.D., for
industry?
assistance in study design. They thank all
• To what extent do clinical researchers en- participating medical schools for their
counter special ethical topics that are not cooperation. They thank Gerard Magill, Ph.D.
covered in general RCR courses? and Dennis Daly, SJ for securing funding for this
These questions remain unanswered. Literature project. This study was funded by the Marchetti
in academic medicine has addressed the roles of Fund at Saint Louis University.
191
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Bibliography
1. Miles SH, Lane LW, Bickel J, Walker RM, Cassel CK.
Medical ethics education. Coming of age. Acad Med
1989;64:705-14.
2. Fox E, Arnold RM, Brody B. Medical ethics education:
past, present and future. Acad Med 1995;70:761-9.
3. Musick DW. Teaching medical ethics. A review of the
literature from North American medical schools with
emphasis on education. Medicine, Health Care, and
Philosophy 1999;2:239-54.
4. Mastroianni AC, Kahn JP. The importance of expanding
current training in the responsible conduct of research.
Acad Med 1998;73:1249-54.
5. Taylor I, Baum M, Cooper A, Johnston IDA. Dilemmas
facing surgical research in the ‘90s. Ann R Coll Surg
Engl 1991;73 Suppl 4:70-2.
6. Friedman PJ. Research ethics: a teaching agenda for
academic medicine. Acad Med 1990;65:32-3.
7. Jonsen AR. Clinical research ethics. Journal of Medical
Education 1975;50: 918-9.
8. Brody B. The ethics of biomedical research. An
international perspective. New York: Oxford University
Press; 1998.
9. Levine RJ. Ethics and regulation of clinical research.
Baltimore: Urban & Schwarzenberg; 1986.
10. Spicker SF. The use of human beings in research: with
special reference to clinical trials. Boston: Kluwer;
1988.
11. Christakis N. Do medical student research subjects need
special protection? IRB: A Review of Human Subject
Research 1985;7:1-4.
12. Makarushka JL, Lally J. Medical schools, clinical
research, and ethical leadership. Journal of Medical
Education 1974;49:411-8.
13. DuBois, JM, Ciesla, JE, Voss, KE. Ethics education in
US medical schools. A study of syllabi. Academic
Medicine. In press.
14. Office of Research Integrity and Public Health Service.
PHS policy on instruction in the responsible conduct of
research. Adopted and published on December 1, 2000.
Available from: URL: http://www.ori.dhhs.gov
15. Cho MK, Shohara R, Schissel A. Policies on faculty
conflicts of interest at US universities. JAMA 2000;
284: 2203-8.
16. Mastroianni AC, Kahn JP. Encouraging accountability in
research: a pilot assessment of training efforts.
Accountability in Research 1999; 7: 85-100.

192
–––––––––––––––––––––––––––––––––––––––––––––––––––––––– DuBois, et al., Research Ethics in US Medical Education

Table I: Rank Order and Content of the Research Ethics Categories

An asterisk * followed by a number indicates that the general category or specific topic overlaps with a PHS Policy “Core
Instructional Area.” The number indicates which of nine instructional areas it overlaps with.
‘Percent valid’ indicates how often a research ethics topic is included in those syllabi from the 23 schools that actually teach
research ethics.
‘Percent all’ indicates how often a research ethics topic is included among all participating schools (i.e., the 87 schools that
returned a survey).

1. CLINICAL TRIALS (*6) – 78% of valid / 21% of all

• Therapeutic vs. non-therapeutic research
• Person as patient vs. research subject
• Physician as clinician vs. physician as scientist
• Selection of subjects for clinical trials
• Randomization
• Patient as research subject vs. health research subject
• Ethics of medical students’ roles in clinical research
• Drug testing and the role of the FDA
• Whether scientific methods provides sole criterion for treatment efficacy
• Industry / university relationships (*possibly 5 & 9)
• Types of clinical trials
2. INFORMED CONSENT (*6) – 70% of valid / 18% of all
• Informed consent in clinical vs. research setting
• Sample consent form for adults
• Emergency waiver of informed consent
• Coercion
• Deception – active and passive
• Placebos
3. GENERAL ETHICS OF HUMAN SUBJECT RESEARCH (*6) – 65% of valid / 17% of all
• Ethics of human experimentation
• Justification of research involving human subject
• Challenges to human subject protections
4. GOVERNMENT COMMITTEES & REGULATIONS (*6 & others) – 61% of valid / 16% of all
• Belmont report
• President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research (1979-
83)
• National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research (1974-78)
[Published Belmont Report]
• Federal regulations
• National Bioethics Advisory Committee
• Declaration of Helsinki
• Practice and regulations
• OPRR reports, Protection of Human Subjects
• Title 45, Code of Federal Regulations, part 46 (1994)
• Nuremberg Code (as living document)
5. HISTORY AND BACKGROUND OF RESEARCH ETHICS – 57% of valid / 15% of all
• Nazi experimentation / Holocaust (awareness of attitudes toward)
• Nuremberg Code (as historical document)
• Tuskegee study of syphilis (awareness and attitudes toward)
• Abuses and errors of early eugenics
• “Frankenstein”
• Sloan-Kettering experiments
• Willowbrook experiments
• Henry Beecher revisited (article by DJ Rothman)
• Introduction to sulfonamides revisited (articles by BH Lerner)
• Research in the Hippocratic Oath (i.e., the fact that it is not addressed therein)

193
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
6. PROTECTING VULNERABLE POPULATIONS (*6) – 52% of valid / 14% of all
• Minorities
• Newborns, Infants, Children
• Soldiers
• Prisoners
• Mentally ill
• AIDS patients
7. IRB (*6) – 48% of valid / 13% of all
• IRB issues
• Definition of research / Novel therapy vs. research
8. RESEARCH INTEGRITY & MISCONDUCT (*8 & 9) – 39% of valid / 10% of all
• Accuracy of published data
• Research fraud (*8)
• Appearance of impropriety
• Scientific misconduct (*8)
• Scientific integrity
• Appropriate credentials
• Research quality guidelines for both academic and non-academic environments
• Conflicts of interest (*9)
9. ETHICAL PRINCIPLES IN HUMAN SUBJECT RESEARCH (*6) – 39% of valid / 10% of all
• Respect autonomy
• Do good (beneficence)
• Fairness / justice
• Avoid harm to subjects (non-maleficence)
• Justify level of risk
• Apply process of ethical decision making to research ethics
10. ANIMAL EXPERIMENTATION (*7) – 30% of valid / 8% of all
• Animal rights
• Use of animals for research
• Poor living conditions for research animals
11. GENETIC RESEARCH AND THERAPY (*6) – 26% of valid / 7% of all
• Genetic research
• Germ-line therapy
• Somatic cell genetic therapy
• National Human Genome Research Institute
• Genetic information and privacy
• Cystic fibrosis research
12. RESEARCH AND THE SOCIAL GOOD (*6) – 22% of valid / 6% of all
• Medicine and the goals of society
• Research in the international context
• Social utility of research
• Relationship between ethics, science, and technology
• Balancing society’s mandates, competing pressures to innovate
13. MINIMIZING RISKS (*6) – 22% of valid / 6% of all
• Establishing gold standard
• Asking whether risk is proportionate to benefit
14. SUBJECT SELECTION (*6) – 13% of valid / 3% of all
• Ensuring the inclusion of women, children and minorities (a concern of justice, rather than protection)
15. EMBRYO AND FETAL RESEARCH (*6) – 9% of valid / 2% of all
• Stem cell research
• Research on live-born fetuses
16. EPIDEMIOLOGY (*6) – 4% of valid / 1% of all
• Ethics of epidemiology
17. MILITARY RESEARCH ETHICS (*6) – 4% of valid / 1% of all
• Experiments related to weaponry
• Using compounds not fully tested in a wartime situation
194
Teaching Ethics in Biomedical Science: Effects on Moral Reasoning Skills
Elizabeth Heitman, University of Texas School of Public Health, Houston, Texas, USA
Patricia J. Salis, University of Texas School of Public Health, Houston, Texas, USA
Ruth Ellen Bulger, Uniformed Services University of the Health Sciences, Bethesda,
Maryland, USA

Key words: Defining Issues Test, DIT, Ethics education, Evaluation, Responsible conduct of research

Academic institutions that train professionals play an important role in ensuring that trainees learn the
ethical norms of their respective disciplines, and that they learn to behave ethically from the start of
their professional lives. The National Institutes of Health requirement that funded research training
programs include education in scientific integrity has made formal courses on the responsible conduct
of research increasingly common in academic medical centers and research universities.
There is still no consensus on what constitutes the most appropriate subject matter, format,
methods, or faculty for teaching the responsible conduct of research. The objectives of general
courses on the responsible conduct of research and scientific integrity typically include increasing
students’ understanding of the norms of scientific practice, their recognition of ethically problematic
situations in science, and their ability to analyze and respond to such situations in a morally mature
manner. Courses vary in the specific content, the number of contact hours, the format (lecture, small-
group discussion, video or web-based tutorials), and the instructors’ professional background and
ethical expertise. The effectiveness of available courses probably also varies. Studies of how students
are affected by formal ethics courses in such disciplines as engineering, law, dentistry, medicine,
nursing, journalism, accounting, veterinary medicine, and social work have found that course design
influences the extent to which students’ ethical reasoning skills change during the courses (1-3). Such
evaluation in the area of scientific integrity, however, is still in its infancy.
The syllabi of courses on the responsible conduct of research in several institutions suggest that
such courses present at least three different kinds of instruction to students. The first is the “how-to”
of science, in which the practical, procedural dimensions of science, rather than its ethical
dimensions, are the focus: how to devise an experiment, give a talk, or write a manuscript. The
second kind of instruction relates to the rules, regulations, and professional norms articulated by the
organizations in which scientists work, their professional societies, and/or the government: how to
make experimental data available for use, how to address suspected research misconduct, and how to
deal ethically with animal and human subjects. Ethical considerations are often addressed as an
aspect of these practical issues. Lecture and individual reading assignments are effective mechanisms
for teaching both of these traditional types of subject matter, and students’ understanding and
retention can be evaluated by an objective written (including computerized) or oral exam.

Corresponding author: Elizabeth Heitman, PhD, University of Mississippi Medical Center, 2500 North State Street, Jackson,
MS 39216, 601-815-1134 (voice), 601-366-3262 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The third type of instruction presented by later continued by Drs. Stanley Reiser and
these courses relates to students’ ability to Elizabeth Heitman, have been to encourage
recognize the ethical aspects of problems that students’ interest in the ethical development and
they encounter in their research, and their ability goals of science, and to teach students to prevent,
to address these issues in a considered way. This recognize, analyze, and resolve ethical conflicts
instruction involves their developing moral in the daily conduct of their work (8).
reasoning skills rather than simply From the beginning, the course has used a
comprehending information, and it frequently combination of formal reading assignments,
uses case discussion or problem-based learning. didactic lecture, and small-group case discussion
Two decades ago the Hastings Center Project on to address a wide variety of issues in the
the Teaching of Ethics proposed three criteria for responsible conduct of research. Its faculty have
evaluating the effectiveness of such instruction: always included both ethicists and bench and
1) whether the student understands the central clinical researchers from various disciplines, both
concepts; 2) whether the student can make as lecturers and as discussion leaders. Most are
cogent oral and written ethical arguments; and senior faculty. Since 1988, the course has been a
3) whether the student can recognize ethical requirement for graduation from the Graduate
problems and examine them rationally (4). This School of Biomedical Sciences, and it is an
evaluation is typically conducted through a more elective for graduate students in the School of
subjective examination using actual case Public Health. For the past four years,
analysis, possibly in a written or oral exam, but approximately 120 students have enrolled in the
ideally in a more interactive setting. course each fall, including 90+ from the
The Hastings Center Project emphasized that Graduate School of Biomedical Sciences’ 22
helping students develop skills to recognize and degree programs, 10-15 students from the School
analyze ethical issues and stimulating their moral of Public Health’s 11 degree programs, and
imagination are fundamental to the effective several post-doctoral fellows from the UT
teaching of ethics. The Association of American Medical School and MD Anderson Cancer
Medical Colleges handbook, Teaching the Center. Students in biomedical sciences typically
Responsible Conduct of Research through a Case take the course in their first semester, while
Study Approach (5), has also stressed the need to others often enroll in the second half of their
enhance students’ ethical awareness and formal graduate study.
problem-solving skills in formal education on the Objective written examinations demonstrated
responsible conduct of research. Ideally, the that the course effectively enhanced students’
courses should have a positive effect on students’ knowledge and understanding of both the
actual and future behavior, helping individuals practical how-to of science and the rules,
avoid ethically problematic behavior and regulations, and professional norms of research
enhancing their ability to resolve unfamiliar that the course addressed. Written analysis in the
ethical conflict appropriately. final exam demonstrated students’ ability to
After several years of teaching a formal identify and consider ethical issues. Students’
course on the responsible conduct of research at course evaluations also confirmed that most of
the University of Texas Health Science Center at them found the course valuable to their
Houston, the course’s organizers sought to assess professional development. However, the faculty
its effects and to determine what outcomes could wanted to assess the more comprehensive effects
be evaluated formally. The course, The Ethical of the course on students’ professional attitudes
Dimensions of the Biomedical Sciences, and behaviors.
originated in 1984 as an institutional response to To affect students’ current behavior and
an incident with a foreign graduate student that shape their future action, instructors of courses in
would have been considered plagiarism for a the responsible conduct of research must have
student schooled in the United States (6, 7). three things: 1) an effective way to teach desired
Consideration of the case highlighted the behaviors; 2) an effective way to motivate
administration’s and faculty’s need to articulate students to adopt these behaviors; and 3) a
the university’s ethical expectations and to teach reliable way to measure behavior change. In a
U.S. academic and professional standards to all broad literature review, we found no clearly
students. The primary objectives of the course identifiable, successful method for teaching
subsequently developed by Dr. Ruth Bulger, and ethical behavior or motivating students to act
196
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Heitman, et al., Teaching Ethics in Biomedical Science
ethically. While there has been work on how best Scores are reported in terms of a P%, which
to evaluate students’ comprehension and measures the extent of “principled” reasoning
retention of information related to ethical behind the individual’s assessment of the cases.
conduct, we found no generally accepted way to Cross-cultural applications have found that DIT
measure the presumed beneficial effect of ethics scores increase consistently with subjects’ age
courses on behavior. and education level.
In the absence of accepted measures of This study explored whether two offerings of
behavior change and future practice, surrogate our course on The Ethical Dimensions of the
measures of the effectiveness of courses on the Biomedical Sciences had an effect on students’
responsible conduct of research are needed. principled moral reasoning, as measured by the
Bebeau (9) and her colleagues have developed a DIT.
set of teaching materials for education in the
responsible conduct of research that considers Methods
four psychological processes in the decision to Following an IRB-approved protocol, a total of
act ethically: moral sensitivity (the ability to 215 graduate students who were enrolled in The
interpret a moral situation and the effects of Ethical Dimensions of the Biomedical Sciences
various courses of action on the parties course were asked to complete the DIT at the
involved); moral reasoning (judgment about beginning (before-course) and the end (after-
which course of action is right); moral course) of the 1997 and 1998 classes. Use of
commitment (intention to do what is right) and individual codes protected students’
moral perseverance (the ability to follow through confidentiality. Computerized scoring by the
with ethical behavior). Their method of University of Minnesota Center for the Study of
evaluating the effectiveness of courses that use Ethical Development generated P% scores.* The
the group’s instructional materials assesses the analyses used students’ change scores — the
essential components of students’ moral after-course test score minus the before-course
discernment and moral reasoning. test score — as the data. A preliminary analysis
Efforts to define, implement, and assess of differences in change scores between the 1997
education in the responsible conduct of research and 1998 classes (t-test, independent samples)
in graduate science programs have parallels in was performed to determine whether it was
medical education, where considerable work has possible to combine the data from the two
been done on the teaching of professional ethics classes. Next the effectiveness of the course in
and the evaluation of such teaching. The effects improving students’ principled judgment by was
of ethics courses on medical students’ moral tested directly analyzing whether their change
reasoning skills have been studied since the late scores differed significantly from zero (t-test,
1970s (10). Such evaluations have linked matched pairs). Finally, an analysis of variance
different types of ethics education with changes (ANOVA) test was run to determine whether
in students’ moral reasoning, and have suggested students’ gender or country of undergraduate
that case-based discussion can significantly education (US or non-US) was related to
increase students’ moral reasoning ability. differential change scores.
The Defining Issues Test (DIT) is the
instrument used most frequently to measure Results
moral reasoning skills and the effects of One hundred seventy-two students (80% of the
education on moral reasoning. The DIT was original 215 students) completed both a before-
developed by James Rest and colleagues at the course and an after-course test, 95 students in
University of Minnesota Center for the Study of 1997 (87% of 109) and 77 in 1998 (73% of 106)
Ethical Development (11). The test is a (Table 1). One or both tests from 14 of these
standardized, computer-scored test that is easily 172 subjects were excluded from analysis based
administered to groups. It is based on Kohlberg’s on scoring criteria used by the University of
theory of cognitive moral development, which Minnesota Center for the Study of Ethical
considers the principle of justice as the highest Development. The final sample therefore
moral good. The DIT presents six morally contained 158 students who had valid scores for
problematic scenarios; the subject ranks the both the before-course and the after-course tests.
importance of various moral criteria for judging Change scores did not differ significantly
how to act, then chooses a course of action. between the 1997 and 1998 classes (t=-0.88,
197
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
p=0.38), so a combined analysis of the two or for the 1998 students alone (Table 5). For the
classes was possible. combined group and the 1997 group, there was
The primary analysis assessed the course’s no significant interaction between the gender
effect on principled judgment: It revealed that factor and the location-of-schooling factor, but
the students showed no significant after-course this interaction was significant in the 1998 group
improvement in principled judgment, as (Table 5). The 1998 data in Figure 2 suggest that
measured by the DIT P% score (Figure 1, Table this result arose from the distinctive pattern
2). Indeed, the pattern in six of the eight sub- among men educated in the U.S. Their after-
groups (Figure 2) was for after-course scores to course scores declined somewhat, while those of
drop slightly. both groups of women and of men not educated
Follow-up analyses of the influence on in the U.S. either improved very slightly or
change scores of students’ gender and location stayed essentially the same.
of undergraduate schooling indicated that neither
gender nor location of education had a significant Conclusions
effect for the combined 1997 and 1998 courses The finding that no significant change had
(Table 3), for the 1997 students alone (Table 4), occurred in P% scores after the course on the

Combined 1997 1998

1997 & 1998 Class Class
Classes
No. people who took at least 1 test 215 109 106

No. people who took 2 tests 172 95 77

( 80 % of 215) (87% of 109) (73% of 106)

No. test pairs sent for scoring 172 95 77

No. test pairs not used 14 11 3
Final no. people or test pairs 158 84 74
(92% of 172) (88% of 95) (96% of 77)
(73% of 215) (77% of 109) (70% of 106)
Table 1. Composition of final study sample.
* one or both tests in pair purged by scorers for invalidity; one of pair purged by us due to absence of valid pair-
mate; scorers failed to process test pair

Group Change t Value * p Value

Score
Mean
(SD)

1997 & 1998 Combined -1.27 -1.3 6 0.17

(11.75)

1997 -2.05 -1.6 1 0.11

(11.64)

1998 -0.40 -0.2 9 0.78

(11.89)
Table 2. Statistical evaluation of course’s effect on DIT P% scores: t-tests
(matched pairs) of change scores (after-course minus before-course) in
combined classes and in each class alone.

198
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Heitman, et al., Teaching Ethics in Biomedical Science

Source of Variance Degrees of Freedom F Value p Value

Gender 1 0.21 0.65
Country of education 1 0.54 0.46
Gender X country interaction 1 0.90 0.34
Error 154
Total 157
Table 3. Statistical evaluation of effect of gender and location-of-schooling on DIT P% scores: analysis of
variance of change scores (after-course minus before-course) in combined classes.

Source of Variance Degrees of Freedom F Value p Value

Gender 1 0.16 0.69
Country of education 1 0.09 0.77
Gender X country interaction 1 0.25 0.62
Error 80
Total 83
Table 4. Statistical evaluation of effect of gender and location-of-schooling on DIT P% scores: Analysis of
variance of change scores (after-course minus before-course) in 1997 class.

Source of Variance Degrees of Freedom F Value p Value

Gender 1 1.75 0.19
County of education 1 0.97 0.33
Gender X country interaction 1 4.86 0.03
Error 70
Total 73
Table 5. Statistical evaluation of effect of gender and location-of-schooling on DIT P% scores: Analysis of
variance of change scores (after-course minus before-course) in 1998 class.

responsible conduct of research was a surprising

and frustrating outcome, given the course’s education in the responsible conduct of research
perceived value within the university and the will require effective means of assessing the
number of studies that report significant changes impact of such programs on students’ knowledge,
in students’ moral reasoning skills after similar awareness, and moral reasoning. Under the most
courses in professional ethics. Even more recent proposal requiring such education in all
perplexing was that students in most sub-groups Public Health Service-funded institutions, a wide
actually showed slight declines in P% scores variety of formats appear to satisfy the new
after the course. credentialing standards. Suggested options range
Upon reflection, the authors concluded that from semester-long academic courses to day-long
principled moral reasoning is only one of a workshops to hour-long web-based tutorials, to
number of skills and concepts that we hope to self-study reading programs. As academic
teach and foster in our course. Much of the research institutions develop the expertise needed
material and related discussion in the course to provide education in the responsible conduct
focuses on common conflicts and practical of research, mechanisms must also be developed
ethical strategies in research and collegial to assess the extent to which these different
interaction. Rest and colleagues (12) noted in formats are effective in enhancing participants’
1999 that Kohlberg’s theories, and thus the DIT, moral reasoning skills. Recent observations
address formal ethical structures of society, what reported by Bebeau and colleagues suggest that
they call macromorality, and do not illuminate some apparently unchanged DIT scores may
the micromoral phenomena of personal, face-to- mask important differences in moral sensitivity
face interactions in everyday life. Thus these and reasoning (13). Expanded use of the DIT
null findings suggest that it is essential to ask should strive to uncover all significant changes in
different questions or use different methods to moral reasoning in order that academic courses
evaluate the complex issue of the outcomes of can target their educational intervention
the course. appropriately.
The establishment and ultimate success of However, if the objective of education in the

199
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
responsible conduct of research is to shape the
behavior of researchers and to reform the culture
of research, methods for evaluating such change
must be developed, and instructors must learn
how to present the rules, regulations, and
professional norms of science in a way that
motivates researchers to adhere to them.

Note
* The Center generated the P% scores using its
new system of validity checks, which should be
considered when comparing these results to those
of older studies.

Bibliography
1. Schlaefli A, Rest JR, Thoma SJ. Does moral education
improve moral judgment? A meta-analysis of
intervention studies using the Defining Issues Test.
Review of Educational Research 1985; 55: 319-20.
2. Self DJ, Baldwin, DC Jr., Wolinsky FD. Evaluation of
teaching medical ethics by an assessment of moral
reasoning. Medical Education 1992; 26: 178-184.
3. Self DJ, Wolinsky FD, Baldwin, DC Jr. The effect of
teaching medical ethics on medical students’ moral
reasoning. Academic Medicine 1989; 64: 755-759.
4. Callahan D, Bok S. Hastings Center project on the
teaching of ethics. In: Callahan D, Bok S, eds., Ethics
teaching in higher education. New York: Plenum Press,
1980.
5. Association of American Medical Colleges. Teaching
the responsible conduct of research through a case study
approach. Washington, DC: AAMC, 1994.
6. Bulger RE, Reiser SJ, Studying science in the context of
ethics. Academic Medicine 1993; (Suppl. 3): 55-59.
7. Reiser SJ, Heitman E. Creating a course on ethics in the
biological sciences. Academic Medicine 1993, 68: 876-
879.
8. Bulger RE, Heitman E, Reiser, SJ. The ethical
dimensions of the biological sciences. New York:
Cambridge University Press, 1993.
9. Bebeau M. Moral reasoning in scientific research. Cases
for teaching and assessment. Bloomington, IN: Poynter
Center for the Study of Ethics and Assessment, 1995.
10. Self DJ, Baldwin DC Jr. Moral reasoning in medicine.
In Moral development in the professions: Psychology
and applied ethics. In: J Rest & D. Navares, eds.,
Hillsdale, NH: Lawrence Erlbaum Associates, 1994.
11. Rest JR. Development in judging moral issues.
Minneapolis, MN: University of Minnesota Press, 1979.
12. Rest J, Narvaez D, Bebeau MJ, Thoma SJ.
Postconventional moral thinking: A neo-Kohlbergian
approach. Mahwah, NJ: Lawrence Earlbaum Associates,
1999: 15.
13. Bebeau MJ. Influencing the moral dimension of
professional practice: Implications for teaching and
assessing for research integrity (abstract). A Research
Conference on Research Integrity, Bethesda. MD,
November 19, 2000.

200
 Figure 1. Mean DIT P% Scores Before and After C
Combined 1997 and 1998 Classes
NORMS*
70
Before course After course < Moral philosophers 65.1
60

47.58 47.13
50
43.93 43.01
< BA graduates 44.85
40.39 36.91 40.04 40.17 < College 43.19
40

201
< Senior high 31.03
30

20 < Junior high 20.00

10
* Data derived with new
validity checks; norms
derived with old validity
0 checks.
Women, US Women, nonUS Men, US Men, nonUS
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Heitman, et al., Teaching Ethics in Biomedical Science
 Figure 2. Mean DIT P% Scores Before and After C
1997 Class and 1998 Class
70 NORMS*

Before course After course < Moral philosophers 65

60

50
50 47 47
44 43 45 44 43 44
43 < BA graduates 45
41 < College 43.19
39
40 38 37 38
36

202
30 < Senior high 31

20 < Junior high 20

10
* Data derived with new
validity checks; norms derived
with old validity checks.
0
US nonUS US nonUS US nonUS US nonUS
Women Men Women Men
1997 1998
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Fostering Research Integrity through Educational Programs: Lessons
Learned at the University of Minnesota
Jeffrey P. Kahn, Center for Bioethics, University of Minnesota, USA
Peggy A. Sundermeyer, Office of the Vice President for Research, University of Minnesota,
USA
Muriel J. Bebeau, Department of Preventive Sciences, University of Minnesota, USA
Virginia S. Seybold, Department of Neuroscience, University of Minnesota, USA

Keywords: RCR curriculum development, Research integrity education programs, Responsible conduct of
research education

The implementation of a Public Health Service (PHS) policy on Instruction in the Responsible
Conduct of Research (RCR) would be a significant challenge to universities because of its broad
inclusion of personnel involved in research. The University of Minnesota is already meeting this
challenge with the delivery of a comprehensive educational program to over 2,000 faculty and
principal investigators (PIs) in calendar year 2000.
The University of Minnesota is a large, land-grant institution. The intellectual diversity of the
institution is reflected in its 21 collegiate units, 3,000 tenure and tenure-track faculty, and 10,000
graduate students enrolled in 150 masters and doctoral programs. The foundation of our educational
programming in RCR developed centrally, early in the 1990’s, to support the educational requirement
of training grants. These programs were expanded to faculty in the mid-90’s in response to growing
institutional and national concern about misconduct in research. The current curriculum is the result
of an institutional corrective action plan initiated by National Institutes of Health (NIH) in 1997.
Therefore, a unique set of circumstances required the University of Minnesota to implement a
comprehensive educational program in RCR before announcement of the PHS policy on Instruction
in RCR.
Our goal is to share the experience of our institution in order to aid others in the development of
programs to meet the requirements of the PHS policy. Points of discussion within the context of the
evolution of the educational program at Minnesota include 1) policy as framework for education,
2) development and delivery of the curriculum, 3) resources and financial investment, and 4)
evaluation.

Policy as Framework in Education

One strength of the educational initiative at the University of Minnesota is that the importance of
RCR is reflected in institutional policies. The Board of Regents, the administrative authority of the

Corresponding author: Jeffrey Kahn, Center for Bioethics, University of Minnesota, N504 Boynton, 410 Church St. SE,
Minneapolis, MN 55455-0346, 612-624-9440 (voice), 612-624-9108 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
University, passed the Code of Conduct in 1996. on meeting the educational requirements.
This policy pertains to all members of the
University community and states that we will Development and Delivery of the
“adhere to the highest ethical standards of Curriculum
professional conduct and integrity.” While The development and delivery of the educational
affirming the common values of research and program in RCR for investigators has been led by
scholarship, it is a clear demonstration of the Faculty Education Advisory (FEA)
institutional ownership of these values. In 1999, Committee. The FEA Committee is in its third
the Board of Regents passed a revised policy on year of existence and is made up of faculty, with
Principal Investigator Eligibility on Sponsored senior administrators serving in ex officio
Projects. This policy requires PIs to complete a capacity. The Committee is staffed by personnel
required education in RCR before any awarded from the Office of the Vice President for
funds are released for spending. The policy was Research. The Committee meets monthly and
implemented March 1, 2001, preceding the PHS has had remarkably consistent participation over
policy by approximately two years and providing the three years. Members were added recently to
the motivation for compliance with the increase representation of disciplines within the
educational requirement. Both policies can be University.
viewed at http://www.ospa.umn.edu/policy/ Members of the FEA Committee are senior
respolcy.htm. and respected faculty and broadly represent the
The University of Minnesota has a strong diversity of the University’s colleges,
tradition in faculty governance, so it is not departments, and programs. The commitment of
surprising that the faculty senate has also faculty leaders, coupled with resources and
promoted RCR. In 1999, the faculty senate commitment from high-level University
passed the policy on Education in Responsible administration, has been crucial to the success of
Conduct of Sponsored Research and Grants the FEA Committee’s effort. The Committee
Management (see http://www1.umn.edu/usenate/ has focused on three areas in RCR education:
policies/grantsmgmt.html). Whereas this policy (1) defining and identifying the target
reiterates the expectation that PIs and project populations; (2) identifying topic areas; and
personnel have the responsibility to behave in (3) implementation.
accordance with the highest ethical standards, it
also defines the responsibility of the University Defining and identifying target populations
to provide individuals involved in research with for RCR education and training
information and resources that support The initial focus of RCR educational
responsible conduct. The policy describes the programming has been PIs, both because it
framework for implementing educational represents the largest group of faculty and staff
programs under the leadership of the Vice responsible for the performance of research, and
President for Research and Dean of the Graduate because the University has a system for
School. It outlines the formation of three certification of PI status. This cohort represented
advisory committees, one for each major nearly 2,000 individuals, from across every
constituency: Academic personnel (including college and a diverse range of departments and
faculty and academic administrators), research research areas.
staff (including graduate and postdoctoral This diversity led to a recognition that
trainees as well as project staff), and education in RCR could not be successful as a
administrative staff (including accounting and “one size fits all” program, and that we needed to
secretarial support). The charge to each of these speak to the needs and interests of researchers
committees is to define the educational needs of from outside biomedical research areas. But in
the constituency, develop the curriculum, spite of the diversity of researchers’ needs, the
recommend delivery formats for the curriculum, FEA Committee agreed on a need to achieve a
propose appropriate recognition/accreditation, shared basic level of understanding for all
and establish appropriate continuing education researchers on a core set of RCR issues. This is
requirements. The Vice President for Research based on the view that all researchers belong to
and Dean of the Graduate School is also charged the University’s research community, and that
with the responsibility of maintaining a database

204
––––––––––––––––––––––––––––––––––––––––––– Kahn, et al., Fostering Research Integrity through Educational Programs
such membership brings with it certain instruction—classroom sessions totaling six
responsibilities, including basic familiarity with hours; web-based instruction for some financial
the rule and issues in areas such as research that and grants management topics, followed by a 1.5
involves human or animal subjects. So while hour classroom session; and in-depth special
many researchers may never engage in human or topic instruction involving a 1.5 hour classroom
animal research, it is unacceptable for them to session, web resources, and case studies.
pass it off as someone else’s problem. For those Because of the number of hours of
researchers engaged in research involving human instruction required and the diversity of
or animal subjects, more in-depth education and investigators who need to participate, a large and
training in those areas is required. In addition to diverse pool of instructors was recruited. We
both basic training for all and in-depth training have between four and six faculty who are
when appropriate, the FEA Committee is prepared to deliver one topic area; faculty are
developing recommendations for continuing paired with relevant professional staff for some
education in RCR. topics. These 37 instructors represent 13
colleges and 3 administrative units, and include 4
Identifying topic areas department heads, and 2 associate deans. While
The FEA Committee’s second task was to all of the faculty agreed to teach in our RCR
identify topic areas for curriculum development. efforts on a volunteer basis, the FEA
Since our efforts pre-dated the PHS/Office of recommended and the University’s Vice
Research Intetrity (ORI) draft of final guidelines, President for Research agreed that formal and
an initial list of topics was drawn from the list of material acknowledgement of their efforts is
suggested topic areas in the ethics requirement appropriate. To that end, funds were committed
for NIH Training Grants (T32). The FEA to provide small professional development
Committee then worked to make the list of topics awards to all faculty participating as instructors
relevant to PIs. The current list of topics in the RCR programs.
includes:
Social Responsibility and Misconduct Resources & Financial Investment
Authorship and Peer Review A cornerstone of our program is faculty
Data Management involvement in the delivery of the curriculum.
Intellectual property Faculty are presenters or facilitators of discussion
for each topic. For some topics they are partnered
Conflict of Interest
with staff who are available to answer more
Fiscal Responsibility technical questions. For example, faculty who
Human Subjects deliver the module on Intellectual Property are
Animal Subjects paired with a staff member from the University
Environmental health and Safety office of Patents and Technology Marketing.
After the PHS/ORI guidelines were issued, Faculty are also involved in revising instructional
we compared our list of topics to the guidelines materials used in workshops and on the web, as
in an effort to assess what changes, if any, are well as the curriculum itself.
needed, and determined that we need to add The commitment of respected, senior faculty,
content on both collaborative science and demonstrated by their leadership on committees
mentoring. or their development of the curriculum, enabled
us to recruit other faculty for the delivery of the
Implementation curriculum. Another critical element for
After identifying the target population, and the recruitment was a detailed syllabus for each topic
topic areas that would be covered, the FEA of the curriculum. The syllabus includes learning
Committee’s last task was to develop strategies objectives, relevant policies, principles, issues for
for implementation. Key components in our discussion, reference materials, and case studies
effort include recruiting instructors with for some topics.
appropriate expertise and experience, drawing One limitation of the curriculum was its bio-
mostly from the ranks of the faculty; and a medical flavor, particularly in case studies,
commitment that face-to-face interaction be part largely because of the disciplines represented on
of the educational experience. the initial faculty advisory committee.
We have employed three separate formats for Recognizing this, we targeted faculty in
205
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
underrepresented disciplines to achieve greater include any compensation for the faculty
balance for delivery of the curriculum. Over 50 involved in the delivery. Based on the average
faculty from 34 departments are currently actual salaries of faculty involved in the
involved in RCR curriculum development or workshops, with an average of 1 – 2 hours
delivery. Besides enriching the curriculum, we depending upon the topic, the value for delivery
believed that faculty involvement throughout the would be an additional $32,300. This does not
University would increase ownership and spread include any estimate of faculty time for
commitment to the RCR. An unexpected preparation or involvement in discussions, via e-
outcome of the diversity of disciplines has been mail or in person, of improvements or additions
the high level of interest maintained by the to the materials, sharing of additional references,
faculty as they see the issues in their topic take or similar and recurring work. Although faculty
on new dimensions and challenges from one were recruited without any hint of monetary
discipline to another. reward, we were able to give those most involved
Besides the demonstrated commitment of small professional development grants of
faculty, a successful educational program in RCR $1,000 – 2,000, for an expense of $24,000.
requires strong support services. Instructional Direct administrative costs include the salary
materials are revised and shared amongst and fringe benefits of 1.75 staff years: one full
presenters. When the faculty audience asks time program coordinator, a 50% administrative
previously unanswered questions, the experts are appointment of a faculty member acting as
consulted. The answers are incorporated into program director; and an administrative fellow
future workshops, and the curriculum and (graduate student). However, the direct cost of
instructional materials are revised as appropriate. additional support services including design and
There are also numerous administrative tasks maintenance of web-based tutorials as well as
associated with scheduling presenters, rooms, registration and recording keeping activities are
and equipment; preparation of printed and web nearly impossible to tally since they are provided
based materials; registration and documentation by a number of centralized offices from the
of attendance; tabulation of evaluations; and graduate school administration to the human
feedback to and coaching of faculty presenters. resources offices.
Although these activities happen mostly behind Hardest yet to calculate are the cost of
the scenes, they are critical to the program. faculty hours spent in participation. Since the
Finally, communication is a critical support University of Minnesota has no formula for
service. Requirements and rationale must be faculty productivity or opportunity costs, one
conveyed to the faculty and other research simple estimate was based on salary. Applying
personnel; progress must be reported to the the composite faculty salaries for Category 1
advisory committee (FEA), faculty senate universities in our region from the March 4,
committees, administrative offices, and academic 2000, issue of Academe and the University of
administrators. All available communications Minnesota fringe benefits rate against the
vehicles are used, including monthly newsletters estimate of 9,600 hours spent by faculty in
of the sponsored projects office and of colleges workshops or reading materials, we estimate the
as well as the University’s multiple publications; cost of faculty participation at $425,000.
printed brochures and flyers; web based home However, the benefit side of this equation is even
pages and events calendars; meeting in person harder to estimate. Certainly the potential
with faculty committees, academic deans, and liabilities exceed the total cost of the program,
special constituencies (IRB); and e-mailings from including loss of faculty time.
the Vice President of Research, Deans, and
Department heads or chairs. Evaluation
So what does all of this cost? The direct
expenses of the 62 workshops for 2,400 Assessing for continuous course improvement
investigators over a 12-month period is the most The RCR curriculum is currently offered in 2
straight forward. Based on actual cost to date for parts of 3 hours each. At the end of each session,
printing of materials, rental of rooms and participants are asked to complete a one-page
equipment, and similar expenses, these direct course evaluation form which asks 1) whether the
expenses are projected to be $48,600, or $15.20 level of content for each topic is appropriate,
per person per session. This amount does not 2) whether the information on each topic is
206
––––––––––––––––––––––––––––––––––––––––––– Kahn, et al., Fostering Research Integrity through Educational Programs
useful, 3) whether the session increased your Documenting faculty participation in an initial
understanding, and 4) whether the materials and and on-going educational program in RCR
resources were helpful. Finally, there is a place demonstrates compliance with a federally
for comments. Data from these forms have been mandated corrective action plan (e.g., the NIH
summarized and used to make course plan currently in effect for the University of
improvements. Minnesota). It does not, however, provide
During the first 6 month period, 66% of the evidence that the attitudes, values, and behaviors
participants (N=522) returned the evaluation for that gave rise to the disciplinary action have
part 1; 43% (N=1162) for part 2. In general, 80% changed. Likewise, installing a model system for
of the participants judged the material presented financial accountability, such as the Electronic
to be appropriate. Lists of resources and Grants Management System (EGMS), can alert
websites were considered the most useful an individual faculty member and his/her unit
resources. Early on, criticisms outpaced head when a proposed action is not within the
satisfactory remarks 3 to 1. Constructive bounds of sanctioned behavior. It does not,
comments included: make the course more however, assure that the moral climate in which
interactive, provide readings ahead of time, research is conducted is enhanced, or will it
incorporate web based materials, and shorten the necessarily improve the ability of investigators to
length of time. Subsequent iterations of the interpret ambiguous situations and identify better
course adopted these suggestions. As a result, choices. If we hope to provide evidence that we
the overall rating of usefulness improved, from have improved the integrity of the researcher and
2.7 to 3.0 on a 4 point scale (with 4 being very climate of the institution, we need measures that
useful) for part 1 and from 2.5 to 2.9 for part 2. assess the more elusive outcomes of the research
In addition, there were fewer critical comments, ethics enterprise and that can be used to examine
and the number of statements of praise increased. the effectiveness of our educational programs
Reflecting on the course evaluation data and and compliance systems.
our efforts at course improvements, we have In Fall of 1999, a faculty group was
identified the following contributors to convened to identify opportunities for assessment
participant satisfaction: of outcomes. The following were identified:
Interactive programming. The more interactive Self-assessment questions in web-based
the program, the more it is viewed as useful. modules. Self assessment items have been
Group size. Smaller groups are better received included in several topics: Fiscal Responsibility,
than larger groups. Intellectual Property, Conflict of Interest,
Informed Consent, Protecting Human Subjects.
Presenters from diverse disciplines. Partici-
Although self assessment items are included, we
pants have been less satisfied when the have decided not to invest resources to assess
presenters are all from the same discipline. knowledge level outcomes.
Topic. Some topics seem to be inherently more University-wide climate surveys to track
interesting than others. For example, perceptions of ethical research practices. The
authorship seems to be rated as most inter- last Faculty and Staff Climate Survey of the
esting irrespective of who presents the University of Minnesota was conducted in 1997,
material. Other topics, like intellectual with a summary reported in 1999. Questions are
property and conflict of interest typically get being prepared for the next survey. The purpose
lower ratings for usefulness. However, when will be to track perceptions of the extent to which
the University climate supports ethical conduct
we have broadened the topic of intellectual
generally. Questions would be directed toward
property to include more on copyright, there ethical research practices as well as issues of
were some improvements in rating. Staff academic integrity.
have speculated that in areas like intellectual Narrative interviews of unit administrators.
property and conflict of interest may be In addition to eliciting their perceptions of the
inherently dissatisfying as it is seldom norms of research conduct, interviews with unit
possible for the presenter to give definitive administrators is a way of identifying areas
answers to questions. needing attention.
Graduate student perceptions of the doctoral
Assessing promotion of responsible conduct experience. Melissa Anderson directs the
207
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Academic Life Project, funded by NSF, which Bibliography
studies the normative experiences of doctoral 1. Rest, J. Moraliaty. In Mussen PH (ser. ed.), Flavell J,
students (see paper by M. Anderson in these Markman E (vol. eds.), Handbook of Child Psychology,
proceedings for additional information on this Cognitive Development, Vol. 3, (4th ed). New York:
study). Wiley; 1983; p. 556-629.
Adaptation of measures of ethical reasoning
and role concept. One reason for the paucity of
information on assessment of instructional effects
in this area is the lack of well-validated outcome
measures. Measures must be grounded in a well-
established theory of ethical development and be
sufficiently user friendly to enable their use for a
variety of purposes. We propose to develop two
outcome measures: (1) a measure of ethical
reasoning and judgment about common problems
arising in the research setting, and (2) a measure
of role concept, i.e., how the researcher
understands his/her role relative to other
researchers. The measures will assess two of the
four dimensions of competence described by
Rest’s Four Component Model of Morality (Rest,
1983). The areas are chosen because prior studies
support the usefulness of the methods for
outcome assessment and for demonstrating the
links between performance and day-to-day
ethical behavior. The two measures will be
modeled after existing measures designed for
assessing the outcomes of ethics education in
dentistry. (See paper by M. Bebeau in these
proceedings for additional information on these
approaches).
In summary, a national effort is required to
design outcome measures that can be used to
assess the effectiveness of institutional education
programs in RCR. Measures must well-
grounded theoretically, well validated, and
sufficiently user friendly to enable their use for a
variety of purposes. Such purposes may include:
1) determining the range of criteria that define
competence in topic areas among different
disciplines, 2) conducting a needs assessment to
identify areas where instructional resources
should placed, 3) identifying individual
differences or problems that require intervention
or remediation, 4) providing feedback to
individuals, departments, and institutions on
research ethics competence, 5) determining the
impact of current programs, and 7) studying the
relationship between competence and ethical
behavior.

208
Being a Scientist: Educating for Ethical Conduct
Chloe Doby Little, Department of Physical Therapy, Western Carolina University, USA
Katherine Le Guin White, Department of Physical Therapy (retired), Western Carolina
University, USA

Key Words: Ethical conduct, Problem-based learning curriculum, Reflection-in-action skills, Reflection-in-
experimentation, RCR

This project is predicated on a reflective way of life for being a scientist as the epistemological
foundation for educating health professions students in the ethical conduct essential for scientific
integrity and progress. Thus, being a scientist exemplifies a reflective way of life; and educating
health professions students for ethical conduct embodies the reflective practitioner epistemology
explicated by Schon in his books, The Reflective Practitioner and Educating the Reflective
Practitioner (1, 2). Schon (1) challenges traditional professional curricula and educators that
continue to implement course content based on the positivist, technical-rational epistemology of the
nineteenth and twentieth centuries. The reflection-in-action epistemology Schon (2) pioneered offers
health professions educators and practitioners a theoretical system of knowledge for helping faculty in
science-based professions education update curricula.
The thesis of this project is that a transitional problem-based learning (PBL) curriculum in the
allied health professions provides an excellent framework for education of reflective practitioners.
Reflective practitioners are problem solvers and ethical scientists. Faculties who are themselves
exemplary reflective researchers and teachers can teach ethics through successful PBL experiences
that guide health professions students in development of ethical conduct as the foundation for their
way of life as science-based, reflective practitioners.
A transitional PBL curriculum in the health professions is structured to guide students from
acquisition of new information and knowledge through application of that knowledge in solving
clinically-based problems to reflection-in-action as practitioners. Put another way, the transitional
PBL curriculum helps health professions students progress from information gathering and
knowledge warehousing to practitioners who know through reflection-in-action and are therefore wise
clinicians rather than master technicians.
Faculties, who are science-based, reflective practitioners and instructors, integrate scientific
research, scholarship, and teaching. Successful implementation of reflection-in-action epistemology
in health professions curricula depends in large measure on the participation of wise, dedicated
faculty whose ethical conduct as scholars and as teachers is manifested in their successful
participation in those reflective dimensions of problem-based learning experiences.

Introduction
Keith-Spiegel, et al., (3) report that scientific misconduct is socialized during undergraduate years
Corresponding author: Chloe D. Little, PW Box 2145, Cullowhee, NC 28723, 828-227-2288 (voice), 828-227-7071 (fax),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
with students believing that significant results responsible conduct in the academic courses.
will earn them better grades. Recent research by This project presents informal and formal
Davidson, et al., (4) lends additional support to methodologies to encourage health professions
these findings. One also can speculate that graduate students to develop reflection-in-action
scientific misconduct reflects the attitudes of skills and values that foster ethical practice in
society. Dishonesty and misrepresentations have health professions services and clinical research.
become commonplace and acceptable in the The ultimate goal is to describe a curriculum for
absence of social sanctions against these promoting active student learning throughout a
behaviors and also as a result of increased series of scientific research courses.
professional competition and increased pressure
to produce. Since the 1940’s the incidence of Implementing Problem-Based Learning
self-reported academic misconduct by college Curriculum in Scientific Research for
students has risen 55-78 percent. (5) Other Graduate Health Professions Students
examples of misconduct include medical school First semester course content includes three case-
faculty applicants misrepresenting research based problems for students to study and discuss:
citations, (6) ethics committees endorsing university-specific guidelines for conduct of
unnecessary research, (7) peer-reviewed journals scientific research, how to construct a research
editors misappropriating authorship, (8) and project, and the virtues of ethical research.
researchers faking data in experiments or failing Second semester course content is focused on
to report unfavorable results (9). Some student implementation of the research project
researchers suggest that there has been a constructed during the first semester. In
“reorientation away from traditional values,” subsequent semesters, students reflectively
especially in scientific inquiry (10) . Others examine with faculty mentors their completed
speculate that fraud and dishonesty in scientific student projects for ethical integrity.
research are the inception rather than the rule Learning issues in the first case-based
(11). problem explored in semester one focused on
Regardless, scientists and institutions must defining scientific misconduct through
maintain quality and integrity in scientific differentiating negligence from deliberate
research if progress and public support are to be dishonesty and examining institutional research
sustained. To promote responsible research, policies, especially distinguishing human and
college and university faculties must sensitize non-human research, confidentiality, and the
future scientists to the critical issues in research obligations of scientific researchers. Students
ethics and guidelines. Also, the National complete an institutional review board proposal
Institutes of Health requirements mandate all for their subsequent projects. The second
institutions participating in training grants show problem progresses students to application of
they provide instructions to faculty and students those skills and behaviors learned in the first
in the principles of scientific integrity (12). case-based problem on the rudiments of
Additionally, the final report of the Commission responsible scientific conduct. Learning issues
on Research Integrity noted the importance of for this case include practicing ethical data
providing “formal and informal educational management and examining the ethical content
opportunities to sensitize both junior and senior of published research studies. The third problem
scientists to critical issues in research ethics and is structured to concentrate student learning on
their institution’s guidelines” (13, p.16). management of conflicting interests,
Although expecting college and university determination of criteria for multiple authorship,
faculties to single-handedly prevent research reporting scientific misconduct, and the process
misconduct is unrealistic, faculties can create by which research grants are awarded.
informal learning environments to promote high Second semester learning issues arise from
standards by engaging students in open reflection on students’ performances as they
discussions of ethical and unethical research begin to conduct their research projects,
practices, carefully supervising and mentoring structured during the first semester. Throughout
student research, encouraging responsible data this course faculty and student reflection-in-
management, and modeling ethical behaviors. action and faculty mentoring become critically
Faculties also can create formal methods for important. Learning experiences during this
integrating the study of scientific values and semester are more informal than those structured
210
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Little & White, Being a Scientist
for the first course. Students complete their solving session and throughout the curriculum.
projects in subsequent semesters, varying from To properly monitor student learning, faculties
one to three. Equally critical throughout these must become proficient in classroom reflective
times is informal student-faculty discussions, behaviors that probe and challenge students’
supervision, and reflection that occurs during thinking conclusions and processes, keep
regularly scheduled small group or individual students involved throughout exploration of the
meetings. problem, adjust levels of challenge to students,
and manage group dynamics so that processes
Benefits of Problem-Based Curriculum move toward constructive resolution of the
and Learning Experiencesfor Faculty and problem. Development of learning materials and
Students writing comprehensive clinical problems that
Students and faculty alike are beneficiaries of challenge students demand faculty creativity and
PBL experiences and curricula. Students develop planning that exceed those faculty demands
problem-solving skills through student-directed imposed by a curriculum predicated on
discussions and information gathering traditional technical-rational epistemology.
assignments. They also learn to become self- Faculties relinquish the resident expert status to
directed and independent learners, habits that become guides for student learning that is
equip them for lifelong learning in practice independent and self-directed. Faculty expertise
communities, even in those remote settings in asking rather than telling, planning and
where colleagues and library resources may be guiding rather than showing is essential for
scarce. As they become more independent successful discussions and problem solving
learners, students begin to actively demonstrate sessions.
increasingly critical, creative thinking.
Assessment of one’s peers during PBL Formal and Informal Methodology
experience is an essential dimension of PBL that Designs
requires active participation of all students in a Problem-based learning methodologies presented
learning group. To that end, students must learn here are designed to encourage first-semester
to assess themselves and their colleagues in health professions graduate students to develop
honest, thorough, deep, and sincere ways. reflection-in-action skills and values for ethical
Learning to work critically in this manner helps practice as clinicians and as researchers. The
students reach greater depths of understanding ultimate goal of the methodology is to promote
the importance of frequently and realistically active student learning in the education of future
evaluating their performance as team members scientists who will consistently demonstrate
and learners; they also become skilled in ethical scientific research behaviors.
applying the same sensitivities to evaluating the As with the previously discussed benefits of
participation and performance of their peers in PBL for students and faculty alike, effective PBL
learning groups. These assessment skills and methodology design occurs only when faculties
values also relate to other aspects of PBL: and students participate successfully in the
information management, creation of measurable process. At a minimum, faculties must openly
knowledge bases for solving problems, and discuss with students during learning group
assessing peers, social and ethical skills, sessions those ethical and unethical behaviors in
communication effectiveness, and the ability to scientific research reported in the literature and in
work effectively as a team member. the faculty member’s experience as a scholar-
Finally, development of leadership skills is researcher. Faculties also must carefully and
fostered through revolving, shared group continuously supervise student research activities
leadership. For each problem-solving session, while mentoring student development as novice
students select a group leader, facilitator, and researchers. To be credible leaders for
recorder. All group members serve in each development of ethical behaviors in students,
capacity throughout a semester. faculties must be personally engaged in ongoing
If PBL is to be successful, faculties must and successful scientific research and
become models and coaches, relinquishing their scholarship.
traditional roles as lecturers and purveyors of Student involvement in design of PBL
information. In this role, faculties develop skills methodology requires full participation of all
that monitor student learning during a problem- group members in researching the literature
211
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
available on ethical and unethical practices in encourages independent learning during which
scientific research. Students also must learn to students develop depth of understanding of
engage faculty and student peers in reflective content (14). Through PBL students become
discussions throughout the problem solving more involved in and responsible for their own
group experiences. learning. The objectives of PBL are to assist the
Finally, students must demonstrate learned process of active learning by students as they
ethical behaviors in their own student research develop effective clinical reasoning skills, such
projects completed after their first semester. as critical appraisal, decision making,
Formal faculty and student responsibilities collaboration, and self-directed learning habits in
for methodology design and successful order to participate effectively and actively in the
implementation are focused on scientifically small group discussions during the problem
rigorous planning and participation guidelines. solving of cases. (15, 16) Each problem should
Faculties are charged with responsibility for be designed to provoke critical inquiry, to
developing curriculum materials that include a encourage independent access to multiple and
series of complex, real world, “ill-structured” diverse learning resources, and to generate lively,
problems to stimulate learning, integration and focused, and pertinent small group discussions.
organization of learned information that ensure Reflection-in-action during and after completion
application of past learning to future problems. of a problem promotes transfer of learning as
Curricular materials include learning objectives well as generation of new concepts (16). Recent
for each PBL problem, definition of PBL research findings suggest PBL curricula are
techniques, appointment of small groups of 5-7 effective methods of learning and that students
student learners, identification and instruction of successfully transfer knowledge and skills in
tutors, guidelines for student leadership process timely and meaningful ways (17, 18, 19).
and responsibilities during group learning Researchers have shown PBL promotes
sessions, and development of assessment tools. higher order thinking skills (16). PBL is a
Beyond these process design matters, the curriculum approach that places students in the
essential faculty responsibility is creating active role of problem solver during the process
multiple cases that form the bases for student of constructing meaning from case-based
learning. Without solid, reality-based clinical problems that mirror real-world situations.
cases, the process cannot proceed as a valid or Throughout the process students develop
effective learning experience. As stated earlier, problem-solving and information gathering
faculty also must model the values promoted as strategies, reflection skills, and discipline-
ethical conduct for scientists. They must specific knowledge bases. In the absence of
consistently demonstrate their ability to reflect- actual clinical experiences during problem
in-action as they participate in the group learning solving discussions, students learn to make
experiences. judgments based on facts, information, logic, and
Students likewise have many formal rationalization alone, they must use higher
responsibilities for achieving successful PBL. thinking orders to justify decisions based on
Students must learn to formulate hypotheses as application of learned principles. Nevertheless,
individuals and as learning team members. They the defining measurement of learning during an
must learn to participate effectively and academic course is the quality of research
responsibly as group members for many produced by the student, an outcome that may
outcomes, including designing a plan to solve the not be evident throughout the span of the course.
problem, researching available and pertinent Therefore, continued supervision and mentoring
information, justifying individual and group of a student’s future research activities beyond
decisions and conclusions, recognizing multiple the first semester is essential for facilitating
acceptable solutions to a given problem, ethical development. The authors believe that
evaluating the performance of themselves, their through PBL students will exhibit reflection-in-
peers, and their tutors, and demonstrating novice experiment skills that will culminate ultimately
reflection-in-action skills and values. in reflection-in-action skills1 as they complete
their student research projects and move toward
Discussion and Conclusion mastery as scientific researchers.
Problem-based learning, based on small group
discussion and clinically-based problems,
212
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Little & White, Being a Scientist

Bibliography 17. Borduas F, Carrier R, Drouin D, Deslauriers D, Trenblay

1. Schön DA. The reflective practitioner: How G. An interactive workshop: An effective means of
professionals think in action. NY, NY: Basic Book, A integrating the Canadian Cardiovascular Society clinical
Division of Harper Collins Publishing; 1984. practice guidelines on congestive heart failure into
2. Schön DA. Educating the reflective practitioner: Canadian family physicians’ practice. Can J Cardiol,
Toward a new design for teaching and learning in the 1998; 14: 911-916.
professions. San Francisco: Jossey-Bass Publishers; 18. Dolmans DH, Schmidt H. The advantages of problem-
1987. based curricula. Postgrad Med J, 1996; 72: 535-538.
3. Keith-Spiegel P, Lee B, Spiegel D, Zinn-Monroe M. 19. Dolmans DH, Gijselaers WH, Schmidt HG, van der
Research described in Scientific Misconduct: An Meers SB. Problem effectiveness in a course using
annotated bibliography by Keith-Spiegel, P., Aronson, problem-based learning. Acad Med, 1993; 68: 207-213.
K., & Bowman, M. 1994. Available from: URL: http://
www.lemyne.edu/OTRP.
4. Davidson EW, Cate HE, Lewis CM, Hunter M. Data
manipulation in the undergraduate laboratory: What are
we teaching? First ORI Research Conference of
Research Integrity; 2000 November 19-21; Bethesda,
MD. 2000.
5. Educational Testing Service. Cheating is a personal foul:
Academic cheating background. 1999. Available from:
URL: http://www.nocheating.org/adcouncil/research/
cheatingbackgrounder.
6. Goe LC, Herrera AM, Mower WR. Misrepresentation
of research citations among medical school faculty
applicants. Acad Med, 1998; 73: 1183-1186.
7. Savulescu J, Chalmers I, Blunt J. Are research ethics
committees behaving unethically? Some suggestions for
improving performance and accountability. BMJ, 1996;
313(7069): 1390-1393.
8. Flanagin A, Carey LA, Fontanarosa PB, Phillips SG,
Pace BP, Lundberg GD, Rennie D. Prevalence of
articles with honorary authors and ghost authors in peer-
reviewed medical journals. JAMA, 1998; 280(3): 222-
224.
9. Friedberg M, Saffran B, Stinson T, Nelson W, Bennett
C. Evaluation of conflict of interest analysis of new
drugs used in oncology. JAMA, 1999; 282(15): 1453-
1457.
10. Bechtel HK, Pearson W. Deviant scientists and
scientific deviance. Deviant Behav, 1985; 6: 237-252.
11. Knight JA. Exploring the compromise of ethical
principles in science. Perspectives Biol Med, 1984; 27:
432-441.
12. National Institute of Health. Requirements for programs
on the responsible conduct of research in National
Research Service Award Institutional Training
Programs. NIH Guide for Grants and Contracts, 1989;
19: 1-35.
13. U. S. Department of Health and Human Services,
Commission on Research Integrity. Integrity and
misconduct in research. Washington, DC: Office of
Research Integrity; 1995.
14. Schmidt HG. Problem-based learning: Rationale and
description. Med Educ, 1983; 17: 11-16.
15. Dolmans DH, Schmidt H. What drives the student in
problem-based learning? Med Educ, 1994; 28: 372-380.
16. Barrows HS. (1998). The essentials of problem-based
learning. J Dent Educ, 1998; 62(9): 630-633.

213
Encouraging Accountability in Research: A Pilot Assessment of Training
Efforts*
Anna C. Mastroianni, School of Law and Institute for Public Health Genetics, University of
Washington, USA

Jeffrey P. Kahn, Center for Bioethics, University of Minnesota, USA

Keywords: Pilot assessment, RCR education and training, Training grants

The objective of this pilot assessment was to describe the response of a sample of grantee institutions
to the federally-mandated training requirement in the responsible conduct of research that is part of
NIH Training Grant (T32) funding. Materials collected by the Department of Health and Human
Services (DHHS) were reviewed and described with the following five research goals:
• describe the target audience for training programs
• describe the locus of instructional responsibility for training programs
• describe whether all trainees at an institution participate in the same training program
• describe the program approaches, materials used and program contents
• create a source of baseline information for planning evaluations of future training programs
• identify areas for further research and analysis

Methods
The sample consisted of a collection of materials assembled by DHHS. These included syllabi, course
outlines, case studies, reading lists, institutional research policies, and other information provided by
training grant recipient institutions about their research ethics programs. In June 1996, the Office of
Science Policy, Office of the Assistant Secretary for Planning and Evaluation, DHHS, sought to create
“a library of course materials that are being used by T32 grantees.” A letter was sent to a stratified
sample of T32 grantees requesting “any training materials currently used to instruct trainees in
research integrity and misconduct” (1). The stated goal of collecting this library of information was
to provide an understanding of training programs in the responsible conduct of research, including the
range of institutional approaches for meeting the training grant requirement. This information was
not collected as part of assessing regulatory compliance or as part of any oversight effort, but to
create a resource and a source of baselines information for planning evaluations of future training
programs (2).1 This sample served as a convenient and best available sample for this review.
Excerpted from: Mastroianni A, Kahn JP. Encouraging accountability in research: A pilot assessment of training efforts.
Accountability in Research 1999;7:85-100. Some policy implications of the results presented here are also discussed in
Mastroianni, A.C. and Kahn, J.P. The importance of expanding current training in the responsibleconduct of research. Acad
Med 1998; 73(12):1249-1254.
Corresponding author: Anna Mastroianni, University of Washington School of Law and Institute for Public Health Genetics
1100 NE Campus Parkway, Seattle, WA 98105-6617, 206-616-3482 (voice), 206-616-3480 (fax),
[email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
DHHS contacted awardees at 50 of the 210 total number of trainees. In examining total
institutions that held training grants as of October number of trainees and number of T32s against
1995. (3) DHHS selected these 50 based on other variables, each was found to be a proxy for
number of training grants, geographical location, the other. Variables, where appropriate, are
status as public or private institution, and number grouped by numbers of T32s only.
of T32 trainees at the institution. For those
institutions with multiple training grants, Results
individual grants were selected for inclusion in There were 45 institutions in the sample
the sample in order to obtain diverse representing 660 T32s (number of T32s at each
representation. Selection factors included: the institution ranges from 1 to 60, with a median of
number of trainees, the distribution of pre- and 6) and 4,883 trainees (number of T32 trainees at
post-doctoral students, and representation of each institution ranges from 3 to 507, with a
clinical and basic research. median of 38). Responses concerning 75 training
DHHS contacted Principal Investigators by grants were represented in the sample.
telephone and follow-up letter, and requested that Of the 45 institutions, 25 [55.6%] were
they provide “any training materials currently public educational institutions, 17 [37.8%] were
used to instruct trainees in research integrity and private educational institutions, and 3 [6.7%]
misconduct [including] materials such as the were non-academic institutions (i.e., a
syllabi, course outlines, case studies, reading professional organization, a non-profit service
lists, institutional codes of conduct in research, provider, and an independent research
etc., [and] any information [that] readily . . . organization).
describes the context in which such materials are
introduced to students and the method of Institutional Characteristics
training” (4). Respondents from 45 of the 50 The sample was reviewed to determine the target
institutions contacted provided information audience for the training programs. Two-thirds of
concerning a total of 75 training grants. institutions represented in the sample required
Access to and copying of these publicly that only T32 trainees receive training in the
available materials was provided by the Office of responsible conduct of research. In this sample,
Science Policy, Office of the Assistant Secretary this result was not affected by the number of
for Planning and Evaluation, DHHS, in training grants held by the institution: 9/14
November 1996. [64.3%] of low-density, 10/15 [66.7%] of
medium-density, and 11/16 [68.8%] of high-
Approach density institutions required training only for T32
A coding form was developed as a method to trainees. Over one-quarter of all of the
collect and summarize information from the institutions, however, required much broader
sample. Descriptive statistics were calculated participation of either all trainees in the school or
using standard statistical software. college, all graduate students or all trainees in the
The characteristics of the sample were institution.
described at the level of either the institution In half (23/45 [51.2%]) of the institutions
(n=45), or the responsible conduct of research represented in the sample, the responsibility for
training program (n=75). In order to understand the responsible conduct of research training
whether institutions shared characteristics based program was located at the departmental or
on number of training grants, the sample of Principal Investigator level. Another quarter
institutions was stratified into thirds by number located the responsibility at the institutional
of training grants. For this purpose, these level. In the materials submitted, 4 [8.9%] of the
groupings were categorized as: “low-density” institutions placed responsibility for the program
institutions (14/45 [31.1%] of the institutions) in their ethics faculty. The institutions that
which held four or fewer training grants; placed responsibility for the program in their
“medium-density” institutions (15/45 [33.3%] of ethics faculty were among the highest-density
the institutions) which held from five through institutions in the sample. They each had 18 or
nine training grants; and, “high-density” more training grants, and represented the top
institutions (16/45 [35.6%] of the institutions) quarter of the sample by number of training
which held ten or more training grants. grants. The majority of low-density and
Institutions also could have been grouped by medium-density institutions had the locus of
216
–––––––––––––––––––––––––––––––––––––––––––––––––––– Mastroianni & Kahn, Encouraging Accountability in Research
program responsibility at the department level two institutions were among the highest-density
[64% and 66%, respectively], while the majority institutions.
of high-density institutions had the locus of Lecture was the most popular method of
program responsibility above the department instruction represented in the sample (53/75
level [75%]. [70.7%]). (Table I) To examine whether
For those 41 institutions with more than one programs relied solely on lectures to satisfy the
NIH training grant, 24 [58.5%] used the same requirement, the frequency of lecture format in
responsible conduct of research program for all combination with other methods of instruction
those required to receive training in the was determined. (Table II) For those programs
responsible conduct of research. As the number that used lectures as a method of instruction, only
of training grants at an institution increased, the a small proportion (4/53 [7.5%]) did not
proportion of institutions utilizing the same supplement lectures with some less didactic
responsible conduct of research training program method or methods of instruction that provide
decreased. Seven of the 10 [70%] low-density, 9 opportunities for greater interaction. It is
of the 15 [60%] medium-density, and 8 of the 16 interesting to note that the materials indicated
[50%] high-density institutions used the same that there was very little use of “brown bag”
program for all trainees. discussions to satisfy the requirement.
Contact hours could be determined for 42 of
Program Characteristics the 75 [56%] programs for which information
The material from the 45 institutions in the was received. The median number of contact
sample included information from 75 training hours for these programs was 10 hours. The
grants. Depending on the characteristic being range was from 4 to 72 contact hours.
examined, the following analyses were based on
either the number of institutions (n=45) or the
number of programs (n=75). The denominator is Method of Instruction* # [%]
noted in each case. Lecture 53 [70.7]
Case study 42 [56.0]
Program approach Small group 36 [48.0]
Submitted materials indicated that one-quarter Seminar 21 [28.0]
of the programs specifically tailored training to Student presentation 11 [14.7]
the trainee population, with either discipline- Mentor 9 [12.0]
specific focus or both general and discipline- Brown bag 1 [1.3]
specific material. Computer 0 [0]
Of the 45 institutions, 28 [62.2%] had a Table 1. Method of program instruction. n=75
formal course in place to satisfy the training * programs could have more than one method of
grant requirement. A greater proportion of instruction
medium-density and high-density institutions
utilized a formal course than did low-density
Methods of Instruction # [%]
institutions: 5 of the 14 [35.7%] low-density Lecture only 4 [7.5]
institutions, 13 of the 15 [86.6%] medium- Lecture + seminar 3 [5.7]
density institutions, and 10 of the 16 [62.5%] Lecture + small group 11 [20.8]
high-density institutions had a formal training Lecture + case studies 16 [30.2]
course in place. Lecture + small group + case 14 [26.4]
Fourteen [31.1%] of the institutions studies
represented in the sample had programs that Lecture + seminar + small 3 [5.7]
indicated the availability of ethics training that group
Lecture + seminar + small 1 [1.9]
could be taken to supplement the course or group + case studies
training offered to satisfy the training grant Lecture + brown bag + small 1 [1.9]
requirement. group
Only two institutions indicated that formal
training was provided to faculty who then carried Table 2.Combination of methods of program
instruction with lectures. Fifty-three programs used
out the required responsible conduct of research
lecture as part of their instructional format. n= 53
training—a “train the trainer” approach. These

217
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Program Contents Rank Content Area # [%]

Material from the 75 training 1 Authorship 65 [86.7]
grants was reviewed to 2 Data Management 56 [74.7]
determine whether course 3 Human Subjects 53 [70.7]
content included the five topic 4 Animal Use 51 [68.0]
areas recommended by NIH in 5 Conflict of Interest 49 [65.3]
the NRSA policy—conflict of 6 Institutional Policy 45 [60.0]
interest, responsible authorship 7 Skills-Writing 44 [58.7]
(including issues of peer-review, 7 Confidentiality 44 [58.7]
plagiarism, and research 9 Skills-Publishing 38 [50.7]
reporting), policies for handling 10 Intellectual Property 37 [49.3]
misconduct (including 11 Mentor/Mentee 35 [46.7]
institutional policies, federal 12 Information Sharing 24 [32.0]
policies, whistleblowing and 13 Whistleblowing and Reporting Misconduct 22 [29.3]
reporting misconduct), policies 14 Moral Reasoning 21 [28.0]
regarding the use of human and 15 Other Content 20 [26.7]
animal subjects, and data 16 Federal Policies 16 [21.3]
management (including 16 Grants Management 16 [21.3]
fabrication, falsification, 18 Skills-Grant Preparation 15 [20.0]
handling research data, 19 Organizational Structure 14 [18.7]
materials and information, and 20 Skills-Oral Presentation 11 [14.7]
data and objectivity). 21 Science and Society 10 [13.3]
Fifty-one [68%] of the T32 22 Laboratory Safety 9 [12.0]
programs covered four or five 23 Skills-Teaching 6 [8.0]
of the NIH recommended 24 Skills-Tenure 4 [5.3]
program content areas while 24 24 Skills-Funding 4 [5.3]
[32%] of the T32 programs 26 Skills-Jobs 3 [4.0]
covered three or fewer of the
categories. The top five ranked
Table 3. Ranking of program content areas. N = 75; programs can have more
categories fell within the five than one content category.
NIH recommended program
content areas, and the top ten ranked categories Thirty-six of the 75 [48%] programs
were addressed by at least half of the T32 provided syllabi or other similar program
programs. (Table 3) materials in the information sent in response to
Content issues that were identified by fewer the DHHS request. Of those, 6 [16.7%]
than half the programs include: identified goals and objectives for the responsible
• whistleblowing and reporting misconduct (22 conduct of research training program. Based on
of 75 programs) this limited information, few programs set forth
• the more theoretical issues encompassed in a traditional goals and objectives for their
category we labeled “moral reasoning” (21 educational efforts.
of 75 programs)
• social issues encompassed in a category we Training Materials
The information submitted was reviewed to
labeled “science and society” (10 of 75
identify the most frequently noted training
programs) materials used by programs. The top
• development of certain skills necessary for three referenced training materials were:
becoming a productive scientist, e.g. grants 1) institutional policies concerning the
preparation and funding, job hunting, oral responsible conduct of research (45/75 [60%]);
communication, tenure, teaching, etc., (3 to 2) Korenman et al., Teaching the Responsible
15 programs). Conduct of Research through a Case Study
General skills related to publishing and writing Approach: A Handbook for Instructors (5) (30/75
received greater attention, with 38 and 44 [40%]); and, 3) the National Academy of
programs addressing them, respectively. Science’s On Being a Scientist (6) (24/75 [32%]).
While the institutional policies are specific to
218
–––––––––––––––––––––––––––––––––––––––––––––––––––– Mastroianni & Kahn, Encouraging Accountability in Research
each institution, Korenman et al. (5) and NAS (6) The finding that a greater proportion of high-
are prepared and widely distributed by density institutions placed program responsibility
professional societies. Of the materials above the departmental level may indicate that as
referenced in the sample, the following four each institutional demand for responsible conduct of
are marketed as offering complete training research training programs increases, more
materials in the responsible conduct of research shared institutional resources are sought.
without need to supplement: Korenman et al. (5); However, based on T32 density, those institutions
Macrina, Scientific Integrity: An Introductory with the highest density had the smallest
Text with Cases (7); the American Association for proportion that utilized the same responsible
the Advancement of Science’s Integrity in conduct of research training program for all
Scientific Research: Five Video Vignettes (8); and trainees. This finding may be attributable to more
Bulger et al., The Ethical Dimensions of the diverse training programs for which different
Biological Sciences. (9) Forty-three [57.3%] of approaches are used, even if some institutional
the programs used one or more of these four resources are shared. Perhaps the administrative
materials. A greater proportion of high-density level at which the ethics training decision is
institutions (12/16 [75%]) used at least one of made affects the institutional approach. Future
these four “ready-to-use” training materials, than research might focus on examining this question,
did low- or medium-density institutions (7/14 and the sharing of institutional resources
[50%]; and 7/15 [46.6%] respectively). regardless of any differences in program
approach.
Discussion The small number of institutions that placed
In this sample, training in the responsible responsibility for teaching in their ethics faculty
conduct of research in response to the NIH may be a reflection of the fact that institutions
requirement was most often directed at T32 with greater numbers of training grants are more
trainees. While the NIH policy encourages likely to have ethics faculty—it would be
expanding training to others, it requires that only interesting to compare the characteristics of
T32 trainees receive such training. If this result institutions that have ethics faculty and place
is representative of institutional commitment to program responsibility in them.
training in the responsible conduct of research, Contrary to the expectations of the authors,
future scientists’ exposure to responsible conduct lecture format alone was rarely used; nearly two-
of research will largely depend on their source of thirds of the programs employed lectures plus
funding. The characteristics of the minority of additional instructional approaches. Also
institutions that make a broader commitment to contrary to popular belief among teachers of
responsible conduct of research education and responsible conduct of research, brown bag
training for its trainees deserve further discussions were rarely identified as an approach
exploration. used to satisfy the training grant requirement.
The T32 recipient institutions in the sample The wide range of contact hours offered by
employed a diversity of approaches to satisfying programs underscores the great diversity in the
the training grant requirement. Approaches implementation of the requirement.
varied both among and within institutions. The majority of programs (51/75 [68%])
Further, the number of T32s held at the specifically addressed four or five of the NIH-
institution had some impact on how the training recommended subject categories. Either the
grant requirement was met. recommendations in the NIH policy have
Locating program responsibility at the influenced program content or the subject
departmental or Principal Investigator level, as categories are well-chosen and represent the
did about half of the institutions in the sample, commonly accepted basic issues in the
may offer ethics training that is more tailored to responsible conduct of research.
the trainees’ disciplines. In the materials Some variation in the subject matter covered
reviewed, a quarter of the programs offered some by programs may result from differences in the
discipline-specific training. Further research is needs of trainees in basic versus clinical research.
necessary to determine whether a relationship However, four of the five NIH-recommended
exists between discipline-specific training and categories are relevant to all scientific research,
location of program responsibility within an i.e, one category, human and animal research
institution. issues, may not be relevant to all researchers.
219
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Therefore, one would expect a higher proportion program evaluation include:
of programs than was observed to address at least • the use of a core program plus tailored
four of the categories. discipline- and skill-specific components
Most educational efforts in other areas • resource utilization-sharing by multiple
typically identify goals and learning objectives as programs within and among institutions
a way of focusing teaching approaches and
• skill-based training programs, with assess-
assessing their success. In this sample, few of the
T32 programs (6/36) identified goals and ment of trainee competencies
objectives. This would seem to imply that • the importance of goals and objectives of
programs do not approach training in the same programs as away to focus the educational
manner they would typically approach other effort
educational efforts. • resource needs for “train the trainer” ap-
“Ready made” materials and materials proaches
sanctioned and made available by professional • the effectiveness of stand-alone training
organizations were a popular source of training materials
materials. This underscores the need to ensure • the effectiveness of one-day programs com-
that these materials, which are so heavily relied
pared to series of sessions
upon, are of high quality, complete, appropriately
There is also a need to identify how to broaden
tailored for the target audiences, and widely
current training efforts to ensure that all
available.
scientists-in-training are prepared to address
The most popularly used material,
ethical dilemmas in their professional careers,
institutional ethics policies, is critical for
regardless of the source of funding for their
trainees’ basic understanding of responsible
training. Such initiatives might include education
conduct of research. The proportion in the
of institutional administrators about the
sample who used these policies as a educational
importance of responsible conduct of research
tool could be viewed as unexpectedly low (45/75
training beyond T32 trainees and the enlisting of
[60%]). institutional commitments for broadened training
efforts. In addition, there is a need for improved
Future Research dissemination of effective approaches to
In addition to the findings discussed, this review responsible conduct of research training in the
indicates the need for further research on relevant professional literature.
institutional approaches to education and training The results of this review should not be
in the responsible conduct of research. First, viewed as representative of responses to the NIH
additional research is needed on the mandate at either the programmatic or
characteristics of training programs. A institutional level because of the sample’s
description of primary and participating limitations. The way the sample was selected
instructors in training would be instructive, and the generality of the government’s request
particularly knowing the extent to which an for materials may have had some impact on the
institution’s ethics faculty are involved in results. Since the materials were collected
program development, administration and independently from this review, a targeted
teaching. In addition, it would be useful to questionnaire would provide more detailed
understand the differences in approach and information. However, the results of this review
content of training provided for trainees in are a valuable first step in describing how
different disciplines, particularly in the clinical institutions and investigators meet the mandate
sciences as compared to the basic sciences. This for training in responsible conduct of research
information would point to differences in the
perceived needs for subgroups of trainees, and
Conclusion
could aid development of appropriate materials
The intent of this pilot assessment was to
and programs, for example, the use of core
describe for the first time how institutions and
programs with tailored components.
investigators are responding to the NIH mandate
Second, research is needed on the
for training in the responsible conduct of research
effectiveness of training initiatives. Evaluation of
that is part of NIH Training Grant (T32) funding.
a variety of programs and their approaches would
The results provide a snapshot of the variety of
be particularly useful. Some target areas for
220
–––––––––––––––––––––––––––––––––––––––––––––––––––– Mastroianni & Kahn, Encouraging Accountability in Research
approaches used in programs across the country. 5. Korenman SG, Shipp AC, AAMC Ad Hoc Committee
Understanding the range of approaches taken on Misconduct and Conflict of Interest in Research
in the education and training in the responsible Subcommittee on Teaching Research Ethics. Teaching
conduct of research is a crucial part of any effort the Responsible Conduct of Research through a Case
Study Approach: A Handbook for Instructors.
to encourage accountability in research, on the Washington, D.C.: Association of American Medical
part of trainees, researchers, institutions, and Colleges, Division of Biomedical Research; 1994.
funders. Those engaged in training and 6. National Academy of Sciences. On Being a Scientist
education can gain important insights for further (2nd Edition). Washington, D.C.: National Academy
study given the diversity of approaches seen in Press; 1995.
this review, while at the same time pointing to 7. Macrina FL. Scientific Integrity: An Introductory Text
the need for some consistency of training with Cases. Washington, DC: ASM Press; 1995.
content. Further, education and training in the 8. American Association for the Advancement of Science.
responsible conduct of research should be part of Integrity in Scientific Research: Five Video Vignettes.
Washington, D.C.: AAAS; 1996.
all the training of all scientists and not a function 9. Bulger RE, Heitman E, Reiser SJ, editors. The Ethical
of the source of funding for training. Only by Dimensions of the Biological Sciences. Cambridge:
assuring the highest standard of research conduct, Cambridge University Press; 1993.
can we be confident that the trust the American
people continue to place in biomedical research
is truly deserved.

Acknowledgements
Thanks to William Raub and Lily Engstrom at
the U.S. Department of Health and Human
Services for providing access to the sample of
materials, and to Wayne Hartley of the Center for
Bioethics at the University of Minnesota for his
assistance in statistical compilation.

Notes
1. DHHS staff selected this approach to the collection of
resources because their primary purpose was to gain
insights into the scope and character of the materials
being used to teach responsible conduct of research, and
in a way that minimized the reporting burden for the
cooperating institutions. They recognized from the
outset that this approach would enable only qualitative
characterization at best, and unlike a formal survey,
would not yield readily analyzable data. (DHHS, 1997)
Bibliography
1. DHHS. Raub WF. Office Communication. Director of
the Office of Science Policy, Office of the Assistant
Secretary for Planning and Evaluation, DHHS, to a
sample of T32 recipient institutions, June 25 [on file
with authors]; 1996a.
2. DHHS. Research integrity project. [Unpublished
memorandum prepared by Preya Sharma on file with
authors]. 1996b Aug 15.
3. DHHS. Raub WF. Personal Communication, DHHS;
1997 Oct 24.
4. National Institutes of Health. NIH program
announcement PAR-98-006: Mentored scientist
development award in research ethics. NIH Guide
Grants Contracts 1997a; 26(37): Nov 7.

221
A Plea for Pursuing New Dimensions of Assessment in the Teaching and
Learning of Research Integrity
Carl Mitcham, Liberal Arts & International Studies, Colorado School of Mines, USA
Barbara M. Olds, Liberal Arts & International Studies, Colorado School of Mines, USA
Ronald L. Miller, Chemical Engineering & Petroleum Refining, Colorado School of Mines,
USA

Keywords: Engineering ethics, Ethics in science, Learning assessment, Responsible conduct of research
education, RCR, Scientific integrity, Science policy, Teaching assessment

Our basic thesis is simple: There are abundant research opportunities involved with the need to assess
the teaching and learning of research integrity. In one sense this thesis is a cliché. Research
opportunities are abundant everywhere; more research can be conducted on almost anything and
everything–even in quite narrowly defined areas such as the quantitative assessment of teaching and
learning about research integrity.
It is nevertheless possible to interpret our thesis in a broader and more provocative sense and to
argue for breaking out of a restricting if well established, four-sided system of constraints. The
teaching and learning of research integrity is, after all, concerned with integrity–from the Latin
integritas, which signifies not only purity or correctness but also and more fundamentally soundness
or completeness, the undiminished or unimpaired wholeness of a thing. Integrity is related to
integritas, bringing together. There is more to ethics than what has been going on in research ethics,
and research ethics will profit from more extensive connections than heretofore pursued.
Before making an effort to move beyond the constraints, it will be useful to describe in slightly
greater detail the two-dimensional box in which this issue of assessing the teaching and learning of
research integrity is currently confined.

Narrow Interpretations of RCR Education

It is increasingly common at research universities to teach courses or modules on research integrity or
the responsible conduct of research (RCR)–as is now required by National Institutes of Health and
Public Health Service grant award guidelines, and as has been reported more generally in Michael
Davis (1). To date, however, efforts to measure the effectiveness of RCR curricula have been limited
if not anecdotal. Nicholas Steneck’s bibliographic background report for the present proceedings
volume begins to identify such limits (2), although he is not as critical as we are of the present state of
affairs.
Constituting a first restriction, the whole literature on research integrity is highly concentrated in
the biomedical field. There are modest exceptions, but the most prominent instances of teaching and

Corresponding author: Carl Mitcham, Liberal Arts & International Studies, Stratton Hall 301, 1005 14th Street, Colorado
School of Mines, Golden, CO 80401, 303-273-3648 (voice), 303-273-3751 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
learning about RCR–and thus of possibilities for modest correlations between ethics education and
RCR assessment–are found in the health care moral reasoning skills, and some possible
fields, from general medicine to dentistry and implications for attitudes or behaviors. Michael
diverse medical research specialities. Given the Kalichman and colleagues at the University of
emphasis on informed consent issues in both California at San Diego (11, 12) have developed
research and clinical practice, and the public an independent instrument that shows similar
profile of regulations related to the treatment of correlations, although other studies (13) raise
animals in research, this is perhaps to be doubts about the full significance of such
expected. It need not, however, be accepted correlations.
without question. No doubt partly as a result of the restrictions
A second restriction is that research ethics in, if not the inconclusiveness of, existing
teaching focuses heavily on what may be termed assessments, it has been argued that the goals of
internalist over externalist issues. Issues ethics education should not be attitudes or
concerned with doing things right crowd out all behaviors at all but simply skills and knowledge
discussions about what might be the right things (14). Indeed, the most common classroom
to do; process overshadows substance. assessments of research ethics teaching
Questions of precisely how to handle data emphasize solely the learning of ethical
management, treat human and animal subjects, reasoning skills, with little attempt to gauge the
pursue publication, deal with conflicts of interest, potential for long-term changes in behavior.
and mentoring protocols dominate, at the expense Arguments have even been made to the effect
of critical reflection on the proper ends to pursue that much more effective than RCR teaching in
with these methods (see the NIH Bioethics the promotion of scientific integrity would be the
Resources on the Web at nih.gov/sigs/bioethics/ establishment of clear behavioral guidelines
researchethics.html, especially the NIH followed by some form of monitoring such as
supported link to Resources for Teaching data audits (15). When education fails, try social
Research Ethics at medicine.ucsd.edu/research/ control.
ethics/resources).
Still a third restriction is that although formal Broader Interpretations of RCR
RCR instruction obviously raises questions about Education
whether such teaching makes a difference– Quantitative assessment of teaching and learning
whether it reduces research misconduct– about research integrity in the academic
confirming evidence remains slight. In fact, classroom is thus boxed in on four sides. Such
there is scant agreement even on the immediate constraint reflects the analytic and reductionist
goals of RCR teaching and learning, thus making strategy of modern scientific methodology, which
it difficult to decide what would count as is based on the demand for and promise of
evidence for or against short- or long-term metrical results; this is a strategy that must
success. In consequence, many assessments of continue to be pursued. At the same time, there
RCR education have produced ambiguous is no need to completely restrict approaches to
results. such a flat plane. Indeed, especially given the
Finally, a fourth restriction is that what wealth of issues associated with moral education,
unambiguous assessment results do exist have there are grounds for stepping beyond such
relied almost exclusively on the utilization and constraints–that is, for expanding our horizons in
adaptation of two specific instruments, the the assessment of the teaching and learning of
Defining Issues Test (DIT) developed by James research integrity.
Rest (3) and the Sociomoral Reflection Measure First, boundaries may be extendend slightly
(SRM) developed by John Gibbs (4), both of by recognizing the limits of particular
whom had studied with, and in their work instruments such as the DIT and SRM. One
attempted to more readily operationalize moral modest movement in this direction would be to
development theorist Lawrence Kohlberg’s consider the relevance of other instruments for
Moral Judgment Interview (MJI). A clutch of assessing cognitive or intellectual development
studies generated by Muriel Beabeau at the such as the Reflective Judgment (RJ) scale
University of Minnesota and her colleagues (5-7) developed by Patricia King and Karen Kitchener
and Donnie Self at Texas A&M University and (16) on the basis of the work of William G.
his colleagues (8-10) all observe measurable if Perry, Jr. (17). It may be noted, for instance, that
224
–––––––––––––––––––––––––––––––––––––––––––––––– Mitcham, et al., A Plea for Pursuing New Dimensions of Assessment
the RJ instrument has been pioneered at the association. Yet lack of assessment is also
Colorado School of Mines (18) and repeated at common among scientific professional societies.
Pennsylvania State University (19) as a tool to Although most societies have codes of ethics that
assess the intellectual development of clearly bear on research integrity, Mark Frankel,
engineering students. Although not focused on director of the Scientific Freedom, Responsibility
moral development, RJ has potential implications and Law Program at the American Association
for ethics learning that deserve exploration. for the Advancement of Science (AAAS), has
Second, precisely because RCR education concluded that few scientific societies are able to
raises research questions about long- as well as tell whether these codes are working (21, 22).
short-term effectiveness, the goals of the teaching Finally, extending such reflection even
and learning about research integrity should further, it reasonably may be argued that such
themselves become themes for research. This internalist issues as data management, the
would constitute, as it were, an initial step off the treatment of human and animal subjects,
flat place of quantitative research. Research into publication protocols, conflicts of interest, and
goals, as opposed to research on the effective mentoring standards cannot in reality be
implementation of goals, calls for more than separated from the focused externalist issues of
quantitative or empirical study. It calls for science and technology policy. Indeed,
historical and philosophical analysis and international recognition of the immoral behavior
reflection. It may be noted, for instance, that of some members of the medical research
current assessment strategies tend to carry establishment during World War II stimulated
forward, more or less uncritically, the applied adoption of the Nuremburg Code for free and
ethics movement that arose during the 1980s. informed consent in human subjects research;
At the very beginning of this revival Daniel political concern in the United States during the
Callahan (20) proposed five goals for the 1980s about the improper behavior of scientists
teaching of ethics in higher education: (a) using public funds has been one of the primary
stimulating the moral imagination, drivers to promote RCR education. Surely both
(b) recognizing ethical issues, (c) eliciting a of these historical points deserve to be taught
sense of moral obligation, (d) developing analytic along with the norms of data management and
skills, and (e) tolerating and reducing peer review.
disagreement and ambiguity. Viewed against the
background of the analytic meta-ethics dominant Three (Intentionally Provocative)
at that time, these were all worthy and even Suggestions
modestly revolutionary goals. Historically, Without attempting to draw definitive
however, the focus has increasingly narrowed to conclusions from this four-fold unsystematic
simply developing analytic skills. The teaching expansion of the RCR educational context, we
and assessment of research ethics has largely would like to pose three summary pleas for the
accepted this narrow inheritance, as is reflected pursuit of new dimensions in assessing the
in the very terminological emphasis on teaching and learning of research integrity. In
“responsible conduct of research.” this way we seek to make common cause with
Philosophically, there are even deeper historical others such as J. Andre (23) who have also called
issues to be raised if RCR education is examined for not limiting professional ethics courses to
in the light of such classic reflections on the moral reasoning analyses.
moral life as those present in the works of Plato, First, in light of the public policy roots of
Aristotle, and Augustine, not to mention the RCR education and the larger philosophical and
Upanishads, the Sutras, the Torah, or the religious traditions of ethics, is it appropriate to
Gospels. focus on reasoning or analytic skills in ways that
Third, reflective reconsideration of the goals slight attitudes and behavior? Would it not be
of teaching and learning about research integrity possible to develop, for instance, an instrument
may stimulate recognition that as much if not for assessing cynicism and idealism among
more pertinent teaching and learning goes on students, and indeed to attempt to counteract a
outside the classroom as well as within it. This too common passive cynicism? Social idealism
recognition may, in turn, promote a search for is an honorable heritage of the scientific
ways to assess meta-classroom learning. One tradition, as exhibited by scientific leaders from
meta-classroom context is the professional Francis Bacon to Albert Einstein. In a talk to
225
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
scientists and engineers at the California Institute university setting. The practical RCR
of Technology in 1931, for instance, Einstein educational potential of student honor codes–
argued that some specific to schools of engineering–perhaps
Concern for man himself and his fate must deserves as much attention as relations to
always form the chief interest of all technical engineering ethics codes.
endeavors . . . in order that the creations of our Finally, does the assessment of teaching and
mind shall be a blessing and not a curse to learning itself not also deserve some assessment.
mankind. Never forget this in the midst of your An assessment of teaching and learning
diagrams and equations (24).
assessment requires both community engagement
Contemporary witnesses to this tradition of
and critical analysis. The practice of any
idealistic science can be found in the public
assessment should be guided by the principles
interest activism of International Pugwash
developed by the Assessment Forum of the
founding member and Nobel Peace Prize winner
American Association for Higher Education (28),
Joseph Rotblat (25) as well as SunMicrosystems
which include the following:
co-founder Bill Joy (26). Introduction to such
moral heros of what may be termed scientific •Assessment is most effective when it reflects
social idealism should not be slighted to carve an understanding of learning as multidimen-
out time for parsing moral dilemmas in conflict sional, integrated, and revealed in perfor-
of interest or authorship adjudication, as mance over time.
important as these may well be. •Assessment works best when the programs it
Second, does research ethics need to be seeks to improve have clear, explicitly stated
conceptualized as distinct from engineering purposes.
ethics, as it has been so far? Does the •Assessment works best when it is ongoing.
engineering/science separation not perpetuate It is our contention that assessing of the teaching
stereotypes of the pure scientist versus the and learning of research integrity has only begun.
applied engineer–images at odds with reality in a This is true not only in the narrow senses
world in which virtually all science is dependent associated with quantitative investigation of
on complex technological instrumentation? RCR, but also in the much broader senses of
Moreover, is it not the case that scientists have attempts to develop relations between RCR and
something to learn from engineers regarding idealistic science activism, engineering ethics
ethics? Long before scientists, engineers and academic codes, and the reiterative
formulated ethics codes at the beginning of the assessment of assessment itself.
20th century; they also began taking them into
the classroom well before scientists (26). Bibliography
In the engineering education community 1. Davis, M. Ethics and the university. New York:
today, considerable attention currently is being Routledge; 1999.
given to ABET Criteria 2000, the new set of 2. Steneck, N. Assessing the integrity of publicly funded
accreditation guidelines developed by the research. Background report. A Research Conference
Accreditation Board for Engineering and on Research Integrity, Bethesda, MD, Nov. 18-20, 2000.
Technology (available at www.abet.org). 3. Rest JR., Narváez D, Bebeau MJ, Thoma S.
Criterion 3, for instance, contains 11 attributes Postconventional moral thinking: A neo-Kohlbergian
approach. Mahwah, NJ: Erlbaum; 1999.
that graduates should possess, including
4. Gibbs, John C., Karen S. Basinger, and Dick Fuller.
“understanding of professional and ethical Moral maturity: Measuring the development of
responsibility.” Many engineering programs are sociomoral reflection. Hillsdale, NJ: Lawrence
developing methods to assess student progress in Erlbaum; 1992.
this area, including the use of such instruments as 5. Bebeau MJ. Designing an outcome-based ethics
the DIT. There are also unexplored possibilities curriculum for professional education: Strategies and
for assessing teaching and learning in evidence of effectiveness. Journal of Moral Education
engineering ethics by correlating results from the 1993; 22(3): 313-26.
Fundamentals of Engineering (FE) and 6. Bebeau MJ, Thoma SJ The impact of a dental ethics
curriculum on moral reasoning. J Dent Educ 1994;
Professional Engineering exams required of all
58(9): 684-92.
professional engineers. 7. Bebeau MJ, Rest JR, Narvaez D. Beyond the promise: A
Research integrity should not be separated perspective on research in moral education. Educational
from academic integrity in the research Researcher 1999 May; 28(4): 18-26.

226
–––––––––––––––––––––––––––––––––––––––––––––––– Mitcham, et al., A Plea for Pursuing New Dimensions of Assessment
8. Self DJ, Wolinsky FD, Baldwin Jr DC. The effect of 27. Mitcham, C. The achievement of “Technology and
teaching medical ethics on medical students= moral Ethics”: A perspective from the United States. In:
reasoning. Acad Med 1989; 64: 755-9. Technology and Ethics. Brussels: In press.
9. Self DJ, Schrader DE, Baldwin Jr DC, Root SK, 28. Assessment Principles. ASEE Prism 1997 Oct; 7(2):
Wolinsky FD, Shadduck JA. Study of the influence of 10.
veterinary medical education on the moral development
of veterinary students. J Am Vet Med Assoc 1991;
198(5): 782-7.
10. Baldwin Jr DC, Daugherty SR, Self DJ. Changes in
moral reasoning during medical school. Acad Med
1991; 66(9) Suppl: S1-S3.
11. Kalichman MW, Friedman PJ. A pilot study of
biomedical trainees= perceptions concerning research
ethics. Acad Med 1992; 67(11): 769-75.
12. Brown S, Kalichman MW. Effects of training in the
responsible conduct of research: A survey of graduate
students in experimental sciences. Science and
Engineering Ethics 1998; 4(4): 487-98.
13. Eastwood S, Derish P, Leash E, Ordway S. Ethical
issues in biomedical research: Perceptions and practices
of postdoctoral research fellows responding to a survey.
Science and Engineering Ethics 1996; 2(1): 89-114.
14. Elliott, D, Stern JE. Evaluating teaching and
students=learning of academic research ethics. Science
and Engineering Ethics 1996; 2(3): 345-66.
15. Shamoo AE, Dunigan CD. Ethics in research. Exp Bio
Med 2000; 224(4): 205-10.
16. King PM, Kitchener KS. Developing Reflective
Judgment. San Francisco: Jossey-Bass; 1994.
17. Perry Jr WG. Forms of Intellectual and Ethical
Development in the College Years: A Scheme. New
York: Holt Rinehard and Winston; 1970.
18. Pavelich, MJ, Moore WS. Measuring the effect of
experiential education using the Perry model. Journal of
Engineering Education October 1996 85(4): 287-92.
19. Palmer B, Marra RM, Wise JC, Litzinger TA. A
longitudinal study of intellectual development of
engineering students: What really counts in our
curriculum? 30th ASEE/IEEE Frontiers in Education
Conference, Proceedings Oct 18-21 2000: S3A-2-6.
20. Callahan D. Goals in the teaching of ethics. In:
Callahan D, Bok S, editors. Ethics Teaching in Higher
Education. New York: Plenum Press; 1980. p.61-80.
21. Brainard, J. 2000. Societies urged to fight research
misconduct. Chronicle of Higher Education April 21,
2000; 46(33): A38.
22. DuMez E. The role and activities of scientific societies
in promoting research integrity. Professional Ethics
Report 2000; 13(3): 1 and 7-8.
23. Andre J. Beyond moral reasoning: A wider view of the
professional ethics course. Teaching Philosophy 1991
Dec; 14(4): 359-73.
24. Einstein A. Einstein on Peace. Nathan O, Norden N,
editors. New York: Simon and Schuster; 1960.
25. Rotblat, J. Taking responsibility. Science 2000 Aug 4;
289: 729.
26. Joy, B. Why the future doesn’t need us. Wired 2000
April; 8.04: 238-62.

227
7. Responsible Conduct of Research Courses
The Responsible Conduct of Animal Research
Lida Anestidou, University of Texas-Graduate School of Biomedical Sciences at Houston,
USA

Elizabeth Heitman, University of Texas-School of Public Health at Houston, USA

Key Words: Animal, Animal research, Curriculum, Responsible conduct of research, RCR

Students in graduate education in the basic sciences have a high probability of using live animals at
some point in their research training. Although animal rights are a volatile issue for public debate, the
use of animals in graduate science education raises little controversy among research trainees. Due to
a National Institutes of Health (NIH) mandate, most graduate science programs today offer
instruction in the responsible conduct of research that may include the ethics of experimentation with
animal subjects1. Similarly, federal requirements for animal research review committees include
provisions for the technical training of students and others conducting procedures with live animals2.
As part of their responsibilities for overseeing the housing and care of research animals and the
safe conduct of research, the veterinary staff of the University of Texas-Health Science Center at
Houston offers formal training sessions in the safe and humane handling of laboratory animals and
proper techniques for a variety of procedures. These sessions are offered regularly and are often
filled well in advance.
The University’s Institutional Animal Care and Use Committee (IACUC) and the veterinarians of
the Center for Laboratory Animal Medicine and Care (CLAMC) are justly proud of their record of
concern for animal welfare and the institution’s humane research practices. Nonetheless, faculty
involved in the required research ethics course at the University of Texas-Graduate School of
Biomedical Sciences at Houston routinely hear comments from first- and second-year students who
feel uncomfortable in their animal work, particularly in mastering routine procedures after the formal
training has ended. Often these comments, made in small group discussions, are about the value of
biomedical research with animals and questions about animal suffering. The same students typically
express unwillingness to ask for help or further instruction for fear of criticism from their faculty and/
or older peers. Nonetheless, many agree that more direct training in the handling and use of specific
research animals would improve their skills, confidence, and attitude toward the work, as well as
improve the quality of their research.
Research in medical education has demonstrated that trainees who ignore or discount their
emotional responses to patients and the pain that medical procedures may cause are at risk of
becoming emotionally stifled, cynical, and even punitive in response to the suffering of others. In
contrast, by including formal attention to the emotional dimensions of patient care, medical educators
have been shown to foster trainees’ compassion and personal satisfaction in their work3. Moreover,
by learning to identify and address their emotional responses directly, medical trainees have been
Corresponding author: Lida Anestidou, UT-Houston Mdical School, Deptartment of Integrative Biology, Pharmacology &
Physiology, 6431 Fannin, Houston, TX 77030, 713-500-5082 (voice), 713-500-7444 (fax),
[email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
found to improve the accuracy of their diagnosis particular attention to the aspects of reading
and treatment. Parallel risks and opportunities assignments, class structure and timing, and the
exist for researchers who use animals, and efforts integration of theoretical material and practical
to address the emotional dimension of animal use skills. One week following the end of the course,
make a valuable addition to the institution’s the instructors and course-coordinators held a
efforts to enhance the integrity of scientific similar debriefing and evaluation session with a
research. special focus on potential changes for subsequent
In response to the perceived need for more course offerings. The following constructive
focused education and hands-on training for suggestions were made by course attendees:
graduate students in the biomedical sciences, the
authors organized a new intensive course entitled Positive points
“The Humane Use of Animals in Biomedical 1. The readings were comprehensive and
Research.” The course offers a highly structured challenging.
and multidisciplinary approach to responsible 2. The practical aspects and methodologic
animal research. Its goal is to provide instruction training were invaluable even to students not
in the ethics and regulatory aspects of animal working in laboratories.
research, approaches to the reduction of the 3. Learning about regulations and IACUC
numbers of animals used in specific protocols, activities from IACUC members was very
including alternative research methods, and enlightening about the practicalities of
extensive practical training tailored to the researchers’ obligations and institutional review.
individual animal model that each participant 4. The information on alternative methods
expects to use. Using a combination of didactic to animal research was important to new
sessions, case discussions, and direct, hands-on researchers considering a variety of techniques.
laboratory instruction under the close supervision 5. The presence, knowledge, and guidance
of institutional veterinarians, the course faculty of veterinarians were a tremendous intellectual
seek to enhance students’ theoretical knowledge and practical asset.
base, technical skills, practical compassion, and 6. The variety of viewpoints presented by
professional confidence. interdisciplinary faculty and guest lectures was
An aspect unique to this course is the useful in understanding the scope of animal
inclusion of structured group discussion intended research and its ethical gray areas.
to help students address their personal 7. Discussion of the personal demands of
experiences, concerns, values, and attitudes research was valuable for integrating
regarding their interaction with animals and the interdisciplinary issues and helpful for students
demands of animal research. Faculty facilitators seeking to come to terms with the demands of
help students recognize and prepare for the their work.
personal and ethical challenges of live animal 8. The intensive class format enhanced
experimentation using a modified version of the rapport among students and faculty.
Balint method, which has been used in medical
education to promote personal awareness and Drawbacks and obstacles
effective, compassionate patient care4. 1. The time commitment in an intensive 2-
The course was offered to graduate students, week format was extremely hard for students to
post-doctoral fellows, research associates and manage along with their regular daily schedules.
technicians across the University for the first 2. The summer offering made scheduling
time in July 2000. The course schedule, faculty assignments difficult because of their
including topics, instructors, and format appears travel schedules and other special commitments.
in Table 1. The list of assigned readings for the 3. The logistical complexity of organizing
course appears in the Appendix. multiple faculty in both classroom and laboratory
was very time consuming for the course
Evaluation (Students’, Instructors’, organizers.
Course Coordinators’) 4. More practical discussion of alternative
As part of the wrap-up on the last day of class, methodologies by practicing researchers was
students were encouraged to provide a needed.
comprehensive evaluation of the course, with 5. Students in science are often
uncomfortable with ethical ambiguity and like
232
––––––––––––––––––––––––––––––––––––––––––––––––––– Anestidou & Heitman, Responsible Conduct of Animal Research
clear answers. a course is variable. Faculty need to identify and
6. Faculty need to focus more on the links address the multiple goals of different students in
between ethical debate, science policy, and different backgrounds throughout the class.
practical demands of research.
7. The costs of laboratory materials for a Conclusion
larger enrollment are likely to be considerable Evaluation by the student and faculty participants
8. Students’ perception of the need for such and a critique of the course by the course

DATE CLASS Topic INSTRUCTOR

Monday Lecture • Historical uses of animals in Heitman

07/17 biomedical research Anestidou
• Ethical and regulatory
perspectives on animals in
biomedical research
Tuesday Lecture • Scientific approaches to refining Heitman
07/18 animal research (the three Rs) Anestidou
• Balint group discussion
Wednesday Lecture • IACUC: its function and Smith
07/19 responsibilities Heitman
• How to fill out animal protocol Anestidou
forms
Thursday Lecture • Alternatives to animal models Heitman
07/20 Anestidou
Bjerckey
Friday Lecture • AAALAC and the Guide Goodwin
07/21 • Housing and standards of care for Blasdel
laboratory animals- Facility tour Heitman
• Balint group discussion
Anestidou

Monday Lecture • Mouse biology, care, and Head

07/24 Lab management
Tuesday Lecture • General anesthesia and pain Smith
07/25 Lab control; rodent-specific protocols;
• Anesthesia matters (video)
• Rodent anesthesia practicum
Wednesday Lecture • Monkey retirement facility Griffin
07/26 Lab speaker Heitman
• Balint group discussion Anestidou
Thursday Lab • Disposition of animals after Blasdel
07/27 research Head
• Euthanasia
Friday Lecture • Wrap up course material Heitman
07/28 Discussion • Evaluation Anestidou

Table 1. The Human Use of Animals in Biomedical Research-Course Outline and Schedule
233
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
coordinators resulted in significant enthusiasm to 2. National Institutes of Health, Alcohol, Drug Abuse, and
repeat it. The course will be offered again in the Mental Health Administration. Reminder and update:
summer 2001 term, using mostly the same requirement for programs on the responsible conduct of
didactic methods and material, but in a less research in National Research Service Award
institutional training programs. NIH Guide 1990; 19
intensive format. The course coordinators, (30): 1.
CLAMC veterinarians, IACUC members, and 3. Gorlin R, Zucker, HD. Physicians’ reactions to patients:
the University’s administration hope that in the a key to teaching humanistic medicine. New England
next few years the course will be developed into Journal of Medicine 1983; 308: 1059-63.
both an integrated part of many students’ 4. Novack DH, Suchman AL, Clark, W Epstein RM,
education at the Graduate School and a Najberg E, Kaplan C. Calibrating the physician.
continuing education course available to Personal awareness and effective patient care. JAMA
researchers and others from outside our 1997; 278: 502-09.
institution.

Bibliography
1. U.S. Department of Health and Human Services. Guide
for the care and use of laboratory animals (revised).
NIH Publication #86-23. Washington, DC: DHHS,
1985.

Appendix.

The Human Use of Animals in Biomedical Research-Course Readings

(by topic)

History of animals in biomedical research; ethical & regulatory perspectives on animals

in biomedical research
• F. Barbara Orlans, “The Beginnings of Institutionalized Animal Experimentation” and “Current Attitudes and Ethical
Arguments” in In the Name of Science: Issues in Responsible Animal Experimentation, New York: Oxford University
Press, 1993: 3-34.
• Caplan, Arthur, “Beastly Conduct: Ethical Issues in Animal Experimentation”, Science, 1983, 406: 159-169
• Brody, Baruch “The Use of Animals in Research” in the Ethics of Biomedical Research: An International Perspective,
New York: Oxford University Press, 1998: 11-30.
• National Association for Biomedical Research, “The Strict Regulations that Govern Research” Animal Research Facts,
http://www.fbresearch.org/research98.htm

Procurement of animals for research and education; Scientific approaches to refining

animal research (the three Rs)
• F. Barbara Orlans, “The Source of Laboratory dogs and Cats: Pound versus Purpose-Bred Animals”, in In the Name of
Science: Issues in Responsible Animal Experimentation, New York: Oxford University Press, 1993, 209-220.
• “Shelter Intake and Euthanasia Trends”, Animal Policy Report 2000, 14 (2): 2.
• Judith Reitman, “From the Leash to the Laboratory”, Atlantic Monthly 2000, 286(1): 17-21.
• “Pet Theft: Urban Myth Makes Useful Propaganda”, FBR Facts (Foundation for Biomedical Research), 2000, 7(2), 2
pages. http://www.fbresearch.org
• Joanna Weiss, “Squid’s Fate: Science of Seafood”, Houston Chronicle (from Boston Globe), June 27, 2000, 3D.
• W.M.S. Russell & R.L. Burch, “Introduction” in The Principles of Humane Experimental Technique. London: Methuen
& Co., 1959, 3-13.
• Alan M. Goldberg, Joanne Zurlow, & Deborah Rudacille, “The Three Rs and Biomedical Research” (editorial), Science,
1996, 272: 1403.
• Ruth Ellen Bulger, “Use of Animals in Experimental Research: A Scientist’s Perspective”, Anatomical Record, 1987,
219: 215-220.
• National Association for Biomedical Research, “Animals in Research 1998”, Animal Research Facts, http://
www.fbresearch.org/research98.htm
• Michael F.W. Festing, et al., “Reducing the Use of Laboratory Animals in Biomedical Research: Problems and Possible
Solutions” — The Report and Recommendations of ECVAM Workshop 29, ATLA 1998, 26: 283-301.

234
––––––––––––––––––––––––––––––––––––––––––––––––––– Anestidou & Heitman, Responsible Conduct of Animal Research

The function and responsibilities of the IACUC

• Robert A. Whitney, Jr., “Animal Care and Use Committees: History and Current National Policies in the United States”,
Laboratory Animal Science 1987, Special Issue, 18-21.
• Henry J. Baker, “Essential Functions of Animal Care and Use Committees”, Laboratory Animal Science 1987, Special
Issue, 30-33.
• “Consensus Recommendations on Effective Institutional Animal Care and Use Committees” Laboratory Animal Science
1987, Special Issue, 11-13
• F. Barbara Orlans, “What Does the Public Have a Right to Know?”, in The Human Use of Animals Biomedical
Research. New York: Oxford University Press: 103-117.
• UT -Houston Health Science Center, Animal Welfare Committee, “Animal Protocol Review Form”, revised 11/1/96.

Alternatives to animal models

• Michael Ballis, “Why is it Proving to be So Difficult to Replace Animal Tests?” Lab Animal 1998, 27 (5): 44-47.
• Richard N. Hill & William S. Stokes, “Validation and RegulatoryAcceptance of Alternatives”, Cambridge Quarterly of
Healthcare Ethics 1999, 8, 73-79.
• Jacques LeClaire & Odile De Silva, “Industry Experience with Alternative Methods”, Toxicology Letters 1998, 102-103:
575-579.
• Seymour Levine & Arthur Saltzman, “An Alternative to Overnight Withholding of Food from Rats”, Contemporary
Topics (American Assn. for Laboratory Animal Science) 1998, 37: 59-60.
• Sharron Kirchain & Robert P. Marini, “A Tissue Harvesting Program as a Method for Implementing the 3Rs of
Biomedical Research”, Lab Animal 1998, 27 (8): 37-39.
• Adrian Smith, Richard Fosse, David Dewhurst, & Karina Smith, “Educational Simulation Models in the Biomedical
Sciences”, ILAR Journal 1997, 38 (2), 82-88.

Standards of care and housing for laboratory animals

• National Research Council, Institute of Laboratory Animal Resources, Guide for the Care and Use of Laboratory
Animals, Washington, DC: National Academy Press, 1996.

Anesthesia and pain control

• Lawrence R. Soma, “Assessment of Animal Pain in Experimental Animals”, Laboratory Animal Science 1987, Special
Issue, 71-74.
• American College of Veterinary Anesthesiologists, “Position Paper on the Treatment of Pain in Animals”, JAVMA 1998,
213(5), 628-630.
• American Association for Laboratory Animal Science, “Position Statement: Recognition and Alleviation of Pain and
Distress in Laboratory Animals”, http://aalas.org
• American Association for Laboratory Animal Science, “Policy #12 — Consideration of Alternatives to Painful/
Distressful Procedures — June 21, 2000. http://aalas.org

Disposition of animals after research

• American Veterinary Medical Association Panel on Euthanasia, “1993 Report of the AVMA Panel on Euthanasia”,
JAVMA 1993, 202: 229-249.
• UT -Houston Health Science Center, Center for Laboratory Animal Medicine and Care”, “Animal Adoption Release
form”, revised 11/1/96.

Wrap up and evaluation

• Diane J. Gaertner, Lele K. Riley, & Dale G. Martin, “Reflections on Future Needs in Research with Animals”, ILAR
Journal 1998, 39: 306-310.

235
An Effective Short Course on Research Integrity
Bernard S. Gerstman, Department of Physics, Florida International University, USA

Key words: Course, Ethics, Integrity, Interactive, Research, Responsible conduct of research, RCR, Training

Ethical conduct in research has always been considered of utmost importance within the research
community. Historically, it was assumed that scientific ethics did not require special training.
Instead, the ethical manner in which to carry out research was presumed to be learned by new
scientists automatically and unconsciously, as if by osmosis, as the technical aspects of the research
were carefully taught by their superiors. This was of course, never true. Mendel and Millikan may
have fudged their data, along with numerous others of less renown.
More recently, consideration has been given to developing methods for training scientists in
research ethics, rather than relying on osmosis (1). Part of the impetus for this change is that the
problems associated with unethical procedures in research have become especially visible to the
public when they occur in research in the health sciences (2). This paper reports on a course of short
duration that is designed to train students efficiently and effectively in the ethical conduct of research.

Design
The course is designed for graduate students and undergraduates who have shown an interest in a
career in science. There is no obvious reason why the course design would not be applicable to
students outside the sciences. At this time, all science majors at the home institution do not take the
course. The science undergraduates who are required to take the course are affiliated with special
programs such as the Research Experience for Undergraduates funded by the NSF as well as NIH
funded programs.
The course is designed to meet for one hour each week and to contain a maximum of 15 students.
If necessary, such as in summer sessions, the course can be compressed into a two-week period, but
some of its effectiveness is lost. This will be discussed later in this section when the reason for this
loss in effectiveness will be clear.
The initial course meetings are organized like a traditional class with the faculty member
explaining various aspects of research integrity and unethical behavior. This is best introduced by a
short (one hour) summary of the general principles of ethics in western society, which can then be
used as the basis for the principles of research integrity and ethics. It is important that this
explanation of ethics in general be presented as a summary. If it is presented in another form, such as
an “Introduction to Western Ethics” or any other form that does not convey immediate de facto
credibility, the course runs the danger of degenerating into a philosophy discussion on ethics in
general. Valuable time will then be taken from the specific goal of training the students in scientific
integrity and the course is likely to be neither short nor effective.
In addition to explaining the principles of research integrity, it also is important to be explicit

Corresponding author: Bernard S. Gerstman, Department of Physics, Florida International University, University Park,
Miami, FL 33199, 305-348-3115 (voice), 305-348-6700 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
about the importance of adhering to these occasionally dismissal. Students initially
principles. Thus, the first few lectures of the consider these penalties too light and suggest
course should cover the following topics: criminal prosecution. The faculty member at this
1) the general principles of Research Integrity point can explain the severe ramifications of
(1); these penalties for the researcher’s career.
2) how scientific progress is enhanced by The third, and last, part of the course is the
adherence to integrity by all researchers; most important for successfully conveying the
principles of research integrity and the necessity
3) how scientific progress is slowed by unethi-
of adhering to these principles. It requires each
cal behavior, or even the perception thereof; student to make a half-hour presentation to the
and class about a case of suspected unethical
4) the direct impact of ethical misconduct in behavior in research that they have investigated
research: through a literature search. The students are
i) wasted money by universities and funding expected to use what they have learned in the
agencies, earlier parts of the course in discussing the
ii) wasted time by researchers who trust the following points:
results of others, and 1) an explanation of what actions constituted
iii) injury or death to patients (biomedical unethical behavior, entailing enough of an
research). explanation of the scientific research so that
The middle part of the course shifts to a other students can understand why the
preceptorial structure with faculty led discussions behavior was unethical;
of selected reading material on recent cases 2) how the unethical behavior was uncovered;
concerning violations of research integrity. 3) what the motivation might have been for the
These case studies summarize the accusations, unethical behavior;
how they were investigated, the decisions that 4) what, if any, penalties (real or intangible)
were reached, and penalties imposed, if any.
were suffered by the perpetrators; and
These case studies can be found in the Annual
Report from the Office of Research Integrity of 5) what penalties the student thinks would have
the Department of Health and Human Services been appropriate.
(3). Information for these presentations can be
These case studies supply concrete examples obtained from books(6,7,8) on the subject,
of the topics discussed in the first part of the science magazines such as Scientific American,
course. The vast majority of cases involve data and with especially well-known and recent cases,
fabrication and falsification. This also presents newspapers and general readership magazines.
the opportunity to discuss types of research Students are informed early in the course about
misconduct that are common but insidious: the presentation and are told to choose a case as
sloppy data taking and self-deception (4). In soon as possible. It is hoped that by giving the
these instances, the researcher is not consciously students several weeks to prepare for their
violating the principles of ethical behavior. presentation, they will use the time to follow a
Unfortunately, because the misconduct is meandering path in their literature search and
unconscious, there is no chance for self- learn about several different cases. If two
correction (5). The case studies are useful in students choose the same case, the second
training the students against sloppy data taking student to notify the faculty member is instructed
and self-deception, which can appear to be, or to pick another case.
easily become, data fabrication or falsification.
The case studies also present concrete Results
examples of a new topic — the penalties suffered The first two parts of the course give the students
by researchers who are found to violate the a customary introduction to the issues of research
principles of research integrity. The usual integrity. The third part of the course is crucially
penalties(3) are disbarring from receiving federal important for consolidating these issues. The
funding for 3 to 5 years, monitoring of a students are enthusiastic about making their
researcher by the home institution, mandatory presentation and peer pressure motivates them to
retraction or correction of publications, and do a thorough job. The presentation forces the

238
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gerstman, Short Course on Research Ethics
students to “step into the mind” of a scientist
who is behaving unethically. This obliges them
to confront the temptations to behave unethically
and solidifies the need for self-vigilance.

Conclusion
A short course can be effective in conveying the
necessity of integrity in research and in training
the students on how to perform research in an
ethical manner. For the course to be effective,
the students must be required to take an active
role. A class presentation by each student is of
crucial importance and the most important
element of the course.

Bibliography
1. Marshall Thomsen, Editor, Ethical Issues in Physics:
Workshop Proceedings, July 17-18, 1993, Eastern
Michigan University, Ypsilanti, Michigan.
2. United States Congress House Committee on
Government Operations: nineteenth report. Are
scientific misconduct and conflicts of interest hazardous
to our health? Washington: U.S. G.P.O., House report /
101st Congress, 2d session ; 101-688, September 10,
1990.
3. Office of Research Integrity (US), Annual Report 1998,
Office of the Secretary, Office of Public Health and
Science, Department of Health and Human Services.
4. Lederberg J, Commentary: Sloppy Research Extracts a
Greater Toll Than Misconduct, The Scientist, February
20, 1995, page 13.
5. Thomsen M and Resnick D. The Effectiveness of the
erratum in avoiding error propagation in physics,
Science and Engineering Ethics, 1995, 1:3.
6. William Broad and Nicholas Wade. Betrayers of the
Truth, Simon and Schuster, New York, 1983.
7. Alexander Kohn, False Prophets: Fraud and Error on
Science and Medicine, Basil Blackwell, New York,
1986.
8. Robert Park, Voodoo Science: The Road from
Foolishness to Fraud, Oxford University Press, New
York, 2000.

239
Resources for Instruction in Responsible Conduct of Research
Michael W. Kalichman, Department of Pathology, University of California San Diego, USA
Francis L. Macrina, Philips Institute of Oral and Craniofacial Molecular Biology,Virginia
Commonwealth University, USA
Jeffrey P. Kahn, Center for Bioethics. University of Minnesota, USA

Key Words: Instruction, Internet, Research ethics, Responsible conduct of research, RCR, World Wide Web

In recent years it has become clear that, despite its importance, training in ethics, standards, and
responsible conduct is too frequently minimal or absent in academic science. This deficit is being
corrected in part by the requirement that fellows funded by National Institutes of Health (NIH)
training grants receive such instruction. This requirement has been important to the development of a
variety of outstanding texts now available (1-8) and a number of very effective, thoughtful programs
developed across the country. However, no network provides ready communication about the goals,
resources, tools, or methods for such programs. As a result, the design and implementation of a new
program in responsible conduct of research (RCR) training can be frustrating if not intimidating.
It can be difficult to pull together material for a new RCR program. Unfortunately, such effort is
frequently duplicated even within the same institution and the resulting RCR instruction is uneven in
quality, topics covered, and audience reached. In addition, it appears that the most likely audience for
these programs has been limited to only those NIH trainees required to take part. This is contrary to
the goal that such training is best met by a program that reaches the broad spectrum of the academic
community including staff, undergraduates, medical students, pre- and post-doctoral fellows, and both
junior and senior faculty. However, with the rapid changes in access to the Internet, the technology is
now available to make formats, examples, contacts, and resources immediately available to any
institution interested in providing effective RCR instruction.
The Internet is now being used for a variety of purposes relevant to RCR instruction (9-17). In
just the last couple of years, these resources have evolved rapidly in both form and content. Many
institutions have created web sites that provide considerable content as well as lists of links to other
sites (9-10), typically in the area of ethics. In addition, many universities now have course materials
posted on the web (11-13) and in some cases Internet-based courses, designed to be run without
traditional classroom meetings (14,15). Finally, web-based information is available on programs such
as the “Survival Skills and Ethics” (16) and “Teaching Research Ethics” (17) workshops for teaching
about the teaching of responsible conduct of research. All of these resources provide important
contributions, but diverse audiences, differences between disciplines, and the frequency of significant
new developments, all minimize the value of any one approach to RCR instruction. The proposed
alternative is a continually evolving web site.

Corresponding author: Michael Kalichman, Ph.D., Director, Research Ethics Program, 0003, University of California, San
Diego, La Jolla, CA 92093-0003, 858-822-2027 (voice), 858-534-4772 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
A web site dedicated to resources on for additional materials. For this review phase,
instruction in the responsible conduct of research the primary goals were to solicit new materials
could provide national access to the most from other institutions, modify the framework of
effective programs, materials, and methods for the site as needed to accommodate the new
such training. The long-term goal would be to resources and reviewer suggestions, annotate the
improve the quality and extent of RCR resources, and publicize the site.
instruction. Such a site would not only make it For evaluation of the web site, reviewers
possible for virtually any institution to develop were asked to rank various aspects of the site’s
an RCR program, but would also increase form and content in a brief online form.
general awareness about what is being done, and Numerical rankings were to be scored using a
what can be done, to enhance instruction in RCR. scale of 1 to 5 (1=very low, 2=low, 3=average,
It is intended that this site would complement, 4=high, 5=very high). Additional questions
not replace, other tools for RCR programs (1-17). asked for specific suggestions to improve the
Given the ongoing NIH requirement for training web site, including recommendations of material
grants to include instruction in RCR and the to be added.
proposed extension of this requirement to all
research staff working on PHS-supported Results
projects, many institutions need help to either The first phase of this project was to develop a
extend limited existing programs or to develop web site framework for presenting resources on
new programs. However, even in the absence of instruction in the responsible conduct of research.
federal requirements, it should be enough to Beginning in September of 1999, work on the
know that access to proven materials and web site began at the University of California,
methods for RCR instruction can only help to San Diego with ongoing assistance from
foster responsibility in the conduct of research. collaborators at Virginia Commonwealth
University and the University of Minnesota.
Methods During the initial months, the web site evolved
The core of the web site was first assembled from through several different formats until a version
materials already available for courses taught at was considered ready for external review. In July
the University of California in San Diego, of 2000, the first phase of external review was
Francis Macrina’s course at Virginia begun. The three planned phases of review were
Commonwealth University and his book on completed by November 1, 2000.
“Scientific Integrity,” and course materials under The first external review was based on a
development at the University of Minnesota. limited release of the web site to four reviewers
The site was initially designed to cover nine (two from government agencies and two from
topic areas: (1) Getting started; (2) Defining the non-governmental organizations). In a series of
goals of an RCR program; (3) Elements of an questions about web site form and content, scores
RCR program; (4) Guidelines, requirements, and averaged between 3.25 and 4.75 with medians
procedures; (5) Resources; (6) Case studies; (7) between 3 and 5. The lowest scores were
RCR programs; (8) Contacts; and (9) Evaluating generally assigned to the appearance and
an RCR program. The plan was that these navigability of the web site. Several valuable
primary divisions would be subdivided into suggestions were made for future improvements,
topics generally considered to be relevant to but one–ease of navigation–was sufficiently
responsible conduct of research (e.g., conflict of important to address before the next phase of
interest, use of animals in research, and criteria review. Based on this concern, the structure of
for authorship). Using this framework for the the web site was considerably modified to
content available in the authors’ institutions, the provide the user with a linear arrangement of
initial goals were to design and implement a topics. This and other changes were completed
framework for the web site, insert materials from by the beginning of August 2000.
the authors’ institutions, and annotate the For a second external review, 13 people were
resources. asked to participate. One of the 13 did not
After completion of the first steps of the respond to the invitation, three declined because
project, the web site was to be improved through of conflicting commitments, but two
an iterative process, including three phases of recommended other choices for reviewers.
external reviews, plus soliciting of suggestions Ultimately, of nine who agreed to review the site,
242
––––––––––––––––––––––––––––––––––– Kalichman et al., Resources for Instruction in the Responsible Conduct of Research

QUESTIONS AVERAGE MEDIAN

1. CONTENT
A. How would you rate the choices of topics covered? 5.0 5.0
B. How would you rate the quality of information provided? 4.6 5.0
2. NAVIGATION
How would you rate the ease for navigating within the Web site? 4.2 5.0
3. APPEARANCE
How would you rate the appearance of the Web site? 3.8 4.0
4. OVERALL
A. How would you rate the likelihood you would recommend this
resource for someone developing a new training program? 4.6 5.0
B. How would you rate the likelihood you would recommend this
resource for someone improving an existing program? 4.8 5.0
Table 1. Second phase of external review (6 reviewers). Using a scale of 1-5 (1 = very low, 5 = very high), the reviewers
answered the following six questions.

three failed to meet the deadline. The six A third external review was begun in
reviewers who responded were from two public September of 2000. A total of 48 people were
universities, one private university, two asked to review the web site by early October; 31
government agencies, and one non-governmental responded that they had the time and would be
organization. willing to do so. Of those, reviews were
A summary of the average and median of the completed by 23 reviewers (16 public
second phase of reviewer evaluations is provided institutions, 4 private institutions, 2 government
in Table 1. The reviewers were extremely agencies, 1 Canadian government agency).
positive about the content of the web site A summary of the average and median of the
(averages of 4.6 to 5.0). Compared to the third phase of reviewer evaluations is provided in
previous round of review, these reviewers were Table 2. Evaluation rankings were generally in
also more positive about navigation (4.2 vs. 3.25- the range of 4 to 5. Lowest scores were for the
3.75). Although considered acceptable, no appearance of the web site (average=3.8) and
reviewer scored appearance of the web site as a highest scores were for the likelihood that the
5. In addition, the reviewers offered many reviewers would recommend this web site as a
practical suggestions for improvements in resource for someone developing a new training
content, navigation, and appearance. program (average=4.8). The reviewers were
again generally positive, but several made

QUESTIONS AVERAGE MEDIAN

1. CONTENT
A. How would you rate the choices of topics covered? 4.4 4.0
B. How would you rate the quality of information provided? 4.1 4.0
2. NAVIGATION
How would you rate the ease for navigating within the Web site? 4.0 4.0
3. APPEARANCE
How would you rate the appearance of the Web site? 3.8 4.0
4. OVERALL
A. How would you rate the likelihood you would recommend this
resource for someone developing a new training program? 4.8 5.0
B. How would you rate the likelihood you would recommend this
resource for someone improving an existing program? 4.5 5.0
Table 2. Third phase of external review (23 reviewers). Using a scale of 1-5 (1 = very low, 5 = very high), the reviewers
answered the following six questions.
243
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
excellent suggestions for changes in structure and “This is a great program. I think its strongest
content to the site. Few comments were repeated feature is the way it brings together a wealth of
across reviewers. Major areas of criticism material in a useful and usable form.”
included:
1. Content: One reviewer was looking for a Based on the reviewer comments, further
pre-packaged RCR course along the lines significant changes were made to the structure of
the web site. As of its release, the structure of the
of the web-based tutorials for training
web site was designed around five topic areas:
researchers working with human subjects. Goals (Goals for RCR instruction), Content
As this reviewer observed, this web site (Suggested RCR topics: Descriptions and reading
does not provide such a course. lists), Format (Formats for RCR instruction:
2. Format: The most frequently voiced Descriptions and examples), Tools (Tools for
concern was that the background (a grid RCR instructors: Texts, cases, and contacts), and
similar to a lab notebook page) was dis- Evaluation (Evaluation of RCR instruction:
tracting. Overview and examples). After checking that the
3. Audience: It wasn’t clear to some reviewers structure of the web site was consistent and that
who the audience (i.e., instructors of RCR all links were active and accurate, the web site
was released for public use on November 1,
courses) was for this web site.
2000.
4. Structure: Several reviewers failed to find
key elements of the web site (e.g., the Discussion
examples of courses) and some pointed out As proposed, a new web site was developed to
confusion about the structure of some of facilitate access to resources for instruction in the
the sections (esp. resources and cases). responsible conduct of research. With the
Related to this problem, a couple of the support of constructive comments from external
links did not work, or did not work as reviewers, an initial version of the web site was
expected. made available to the research community
beginning on November 1, 2000. Based on
Several of the reviewers were quite supportive of reviewer comments, this web site will be of value
the web site, for example: both to those first developing programs of RCR
“The choice of topics to be covered in teaching instruction and also to those seeking to improve
research ethics is excellent. I particularly think on existing programs of instruction.
it is useful that ‘minimal instruction’ is defined To achieve the long-term goals for this web
for each of the topics and that more advanced site, it will be necessary for the site to evolve
versions of the units are also suggested. This both in terms of content and format. For this
will be quite helpful to faculty who are just purpose, the authors intend to solicit the latest
beginning to teach RCR, and who want to know information about content and format of existing
what is the minimum level of instruction they RCR programs nationwide. Further, it will be
need to meet.” important to include mechanisms for ongoing
evaluation of the merits of the resources listed on
“I think the site looks great. It is very well
the web site and the web site itself. During this
organized. It will be especially useful for
newcomers.” next phase, the primary goals will be to survey
existing programs in RCR, solicit new materials
“Best collection of materials related to RCR I from these institutions, continue modifying the
have found. The logical progression of steps framework of the site as needed to accommodate
should make it easy to develop or improve the new resources, and implement mechanisms
courses without becoming overwhelmed by the for evaluating the effectiveness of the web site
task at hand. Linked pages were relevant and and the resources listed.
provide materials for inspiration and contrast.”
Acknowledgments
“This is a very strong site and I learned a lot The Office of Research Integrity (ORI) and
just skimming. The links for case studies and particularly Alicia Dustira, Lawrence Rhoades,
analysis of instructional delivery options were
and Chris Pascal, all of ORI, are thanked for their
quite good.”
encouragement and support for the initial
244
––––––––––––––––––––––––––––––––––– Kalichman et al., Resources for Instruction in the Responsible Conduct of Research
development of this web site. The University of
California, San Diego is thanked for its support
of the Project Director (MWK), helping to make
this project possible. In addition, the generous
contributions of time and suggestions of the 33
external reviewers are gratefully acknowledged.
This work was supported by a contract from the
ORI to UCSD (contract #99T07035001D)

Bibliography
1. Bulger RE, Heitman E, Reiser SJ. The Ethical
Dimensions of the Biological Sciences. New York:
Cambridge University Press; 1993.
2. Elliott D, Stern JE. Research Ethics: A Reader. Hanover,
New Hampshire: University Press of New England;
1997.
3. Grinnell F. The Scientific Attitude. Second edition. New
York: Guilford Press; 1992.
4. Korenman SG, Shipp AC. Teaching the Responsible
Conduct of Research through a Case Study Approach: A
Handbook for Instructors. Washington, D.C.:
Association of American Medical Colleges; 1994.
5. Macrina FL. Scientific Integrity: An Introductory Text
with Cases. 2nd edition, Washington, D.C.: American
Society for Microbiology Press; 2000.
6. National Academy of Sciences, National Academy of
Engineering, Institute of Medicine. On Being a
Scientist: Responsible Conduct in Research.
Washington, D.C.: National Academy Press; 1995 [also
available online at: http://www.nap.edu/readingroom/
books/obas]
7. Penslar RL. Research Ethics: Cases & Materials.
Bloomington, Indiana: Indiana University Press; 1995.
8. Stern JE, Elliott D. The Ethics of Scientific Research: A
Guidebook for Course Development. Hanover, New
Hampshire: University Press of New England; 1997.
9. Center for Bioethics, University of Pennsylvania: http://
bioethics.net
10. Online Ethics Center for Engineering and Science: http:/
/onlineethics.org
11. Research Ethics Program, University of California, San
Diego: http://ethics.ucsd.edu
12. Scientific Integrity, a one-semester course, Virginia
Commonwealth University: http://www.vcu.edu/
courses/rcr
13. Teaching Ethics for Research, Scholarship, & Practice,
University of Minnesota: http://www.research.umn.edu/
ethics
14. Research Ethics, an Internet-based course, University of
Nebraska Medical Center: http://www.unmc.edu/ethics
15. Scientific Integrity, an Internet-based course, University
of California, San Diego: http://ethics.ucsd.edu/courses/
integrity
16. Survival Skills and Ethics, University of Pittsburgh:
http://www.edc.gsph.pitt.edu/survival
17. Teaching Research Ethics, Indiana University: http://
www.indiana.edu/~poynter/tre.html

245
An Interactive Web Site for Ethics Training: http://storiesandquestions.com
Rudolph J. Marcus, Stories and Questions, Occidental, CA, USA

Keywords: Case studies, Computer-based instruction, Ethics training instrument, Moral advisor, Self-directed
learning, Work environment

This paper reports on the construction and use of a web site for ethics training. The paper is divided
into three parts. Various uses of the web site as a delivery vehicle for ethics training are outlined in
the first part. Practical advantages of an ethics training instrument usable by individuals at their own
pace, place, and time are discussed in the second part. The web site itself and its operation are
described in the third part located after the references. The paper ends with suggestions for further
work in adding more seminars to the web site, further measuring the web site’s effectiveness, and
developing guidelines for facilitators.

Ethics Training With The Web Site

The computer-based ethics training instrument is conceived as a delivery Types
technique. What might such a training instrument contain? Codes
After the apprentice model lost its effectiveness for ethics training it was Case studies
replaced by a recital of the appropriate ethics codes (Table I). Discussion of case
Awareness
studies is used for ethics training that is more specialized and pointed toward
Table I: Types of
particular disciplines and tasks. Both training in ethics codes and in case studies ethics training
can be delivered by this sequenced text-and-question technique. The present web
site adds a third category to the ethics training instrument, awareness training.
Experience with the web site has shown different results from bulk presentation and from
sequenced presentation. Bulk presentation,
Training Message where the whole story and all of the questions
are presented at one time, usually draws either
What are the limits? no response or a “head trip” response. The
Codes sequenced presentation of a part of the story at a
What can I get away with?
time or an exercise with the story, accompanied
What are the limits? by a single question, appears to encourage the
Case studies thoughtfulness and inner work that lead to real
What can I get away with? attitude change. It is that experience that leads
to the statement of the previous paragraph that
What is going on here? the necessary ethics training of the ethics codes
Awareness What can I do about it? themselves, and of applicable case studies, can
be delivered by this sequenced text-and-question
What might be the right thing to do?
Table II Messages from various types of ethics training
Corresponding author: Rudolph Marcus, 17010 Burl Lane, Occidental, CA 95465, 707-876-1959 (voice),
[email protected]
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
technique. Practical Advantages Of Web Site Ethics
The implication that trainees get from recital Training
of appropriate ethics codes or from discussion of
case studies is “What are the limits?” The Self-directed learning: No travel or time off
message conveyed by such training is “What can from work
I get away with?” (Table II). The web site is an alternative learning mode for
By contrast, awareness deals with questions ethics training which permits study and testing at
such as “What is going on here?” and “What can each individual’s time, place, and pace. It
I do about it?” The message conveyed by reduces or eliminates trying to get everyone
awareness training is “What might be the right together at a given time and place, with resultant
thing to do?” schedule changes and resistance to “a waste of
The reader may sense that I find the “What time.”
might be the right thing to do?” more significant “Valuing of self-directed learning” (2)
and necessary in ethics training because it goes applies to academic institutions and to needs for
beyond the given guidelines. However, ethics training there as well as in industrial and
knowledge of applicable ethics codes and of their governmental laboratories. Gunzburger writes.
application is case studies is an essential and “The survey results indicate that most schools
equally necessary part of ethics training. I have have not established systems for the valuing of
discussed the multiplicity of the overlapping curriculum time that include balancing values for
ethics codes that researchers are subject to and lecture time and values for self-directed
some of the conflicts between those codes in an learning.”
encyclopedia chapter (1). One surprise of experience with the web site
Results of ethics training by recital of was the amount of motivation which is
appropriate ethics codes usually are measured by generated by the sequenced self-study. No
attendance at the training, or by recall of the coaxing or arguments by the facilitator were
Training Results reported as needed to get clients to complete a seminar. If
there was no response to a session within two
Codes Attendance or recall
weeks, re-presentation of the segment was
Case studies Discussion of case sufficient to elicit a response, often with an
Awareness Attitude changes explanation of what the holdup had been.
Table III: Results from various types of ethics
training reported as Engendering an ethical environment in self
and in work place
contents (Table III).
Results of ethics training with case studies An unspoken assumption seems to be that
are measured by qualities of participants’ infractions of an ethics code are deliberate; if not
discussion of the cases. Results of awareness deliberate, then the incident was an “accident”
training can be assessed by noting attitude (3). Based on that assumption, an infraction can
changes. The same question is asked near the be stopped and the perpetrator exposed and/or
beginning and near the end of each seminar. In punished.
experience so far with the web site, and in the Inherent conflicts and inconsistencies
talks and seminars on which the web site is between the different kinds of ethics codes, as
modeled, the response after the second time the well as in individual codes themselves, are
question is asked is much better elaborated, is discussed in the encyclopedia article previously
more grounded in the story, and shows a better referred to (1). There are few, if any, places
grasp of how the story applies to the participant’s where those caught in a conflict between
own life, problems, and actions. Changes different kinds of codes can get help and advice.
become apparent in the perceived self image of The encyclopedia article concludes with a
the seminar participant. The changes are training program that I have designed for
reinforced by the facilitator calling attention to situations like that, and this training program has
them or asking whether the participant has now been put on the web site described in this
noticed the changes. paper. In brief, the seminar participant who uses
the web site ethics training learns where to look
for the advice needed to cope with overlapping

248
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Marcus, An Interactive Web Site for Ethics Training
and conflicting ethics codes to alleged scientific misconduct and how to deal
A different way of stating the problem is with them creatively.
found in a letter to a college student who asked Material is at hand for expansion of the web
for advice on ethical behavior In the letter, C. G. site to about 12 seminars during the coming year.
Jung describes codes of moral values as “general The immediate next additions to the web site will
and not specific,” and therefore “they don’t be four seminars dealing with the origins of
exactly apply to individual situations.” Jung’s science (9). Together, they show four successive
advice to the student is “try to live as stages of scientists working “consciously,
consciously, as conscientiously, and as conscientiously and as fully as possible.” I have
completely as possible (italics are Jung’s) and used that material about twenty times in national
learn who you are and who or what it is that tour lectures for the American Chemical Society
ultimately decides” (4). under the title of “Nature and Gods, Science and
The problem also has been stated in a law Scientists.”
review article about ethics in government, which Further research aspects of this work consist
applies equally well to ethics in science (5): of:
“Our current obsession with investigating and 1. Adding more seminars to the web site.
prosecuting individual wrongdoing may actually 2. Assessing its effectiveness. Effectiveness
prove counterproductive in our efforts to promote can be gauged by looking at changed opin-
ethical [science], promoting instead public ions, feelings, or assessments of problem
cynicism about [science]. To counteract this
situations by seminar participants as the
emphasis on individual wrongdoing, we need to
pay more attention to ensuring that [science] seminar progresses. Records of such
institutions are designed to engender an ethical changes are already being kept while main-
environment.” taining seminar participants’ analytical
All three ways of stating the problem confidentiality, which is a hallmark of the
contradict the beginning assumptin—the old seminars in workshop and web site modes.
vocabulary—that infractions of ethics codes are 3. Developing guidelines for facilitators. As
deliberate or an “accident.” All three statements more people use this method of self-study
indicate that infractions of ethics codes are NOT for ethics training, they too may want to
always deliberate and that the perpetrators may become facilitators and learn more by
not even be aware of their inadvertent and often
helping others to start self-study in scientific
avoidable errors affecting research integrity.
Jung’s advice to “learn who you are” to ethics.
behave ethically is exactly the aim of the training
program that I described in the encyclopedia Bibliography
1. Marcus RJ. Government funding of research. In:
article. The training is to live with the opposites
Encyclopedia of Applied Ethics. San Diego: Academic
in resolving conflicts of various ethics codes Press; 1998. Vol. 2, p. 471-80.
within which researchers have to work (1, 6). It 2. Gunzburger LK. U. S. Medical Schools’ Valuing of
is that kind of training program that I have now Curriculum Time: Self-Directed Learning versus
brought out of the lecture and workshop stage Lectures. Academic Medicine 1993; 68: 700-702.
and put into the web site for one-on-one work 3. Marcus RJ. Ethical considerations. The New Yorker
with an experienced ethics consultant. It is a 1996 March 11; p. 14.
self-contained course that meets a previously 4. Jung CG. Letter to William Kinney (1956 May 26). In:
unrecognized need. Letters. Princeton NJ: Princeton University Press;
1973. Vol. 2, p. 300-1. .
5. Clark K. Toward More Ethical Government: An
Present Status And Future Work Inspector General for the White House. Mercer Law
At this time the web site contains four such Review 1998; 49: 553-64.
seminars. One of the seminars deals with 6. Marcus RJ. Ethics in research and government funding.
collegiality and civility in the work place or, to Chemtech 1997 Nov, p. 8-13.
see it from the other side, conflict resolution in 7. Marcus RJ. Ethics workshops when collegiality fails.
highly polarized situations (7). Another deals In: Abstracts of Annual Meeting, Association for
with how to find a moral advisor in a Practical and Professional Ethics. 1998; Paper V H.
hierarchically structured work environment (8).
Both describe work environments that often lead
249
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
8. Marcus RJ. Sequenced self-study for constellating a discussion group, and there is not a consensus to
moral advisor. In: Abstracts ofJoint Services be reached. Rather, each response is respected as
Conference on Professional Ethics (JSCOPE 2000). that person’s truth at that particular time and
9. Marcus RJ. Mythological stories as case studies for place. In such a workshop, there would be a long
teaching scientific ethics. In: Abstracts of 1996 Annual
Meeting, Association for Practical and Professional
break after the discussion of the material in
Ethics. 1996; Paper II H, Session I. That break might even take the form of
lunch, a nap, a walk in the woods, and/or a swim.
More thoughts about the story, and additional
The Web Site responses to the questions occur, and those might
be written in a journal or one’s workshop notes.

Introduction Stories
Welcome to this web site, the home of “Stories In a group seminar using this material, the
and Questions,” a personal journey in self facilitator would have warned participants NOT
enrichment. Here is a method of exploring who to identify with any of the characters in the story.
one is by reading short stories and responding in That is important and it applies as well to the
the moment to simple questions. These stories self-study.
allow one to stop and feel, and the question The seminar participant encounters the story
permits feelings about one’s life and its direction. as if the participant were seeing it on a stage. The
Stories and Questions is a series of individual participant is not on the stage with the story
seminars facilitated by Rudy Marcus. Rudy has characters. The participant is in the audience
done stories and questions for 16 years (and if watching the actions of all the characters, being
one counts his research experience in the privy to the knowledge, habits, and actions of all
sciences, for 50 years) and has experienced for the characters at that point in the story.
himself their ability to effect personal change. In the language of psychology, the
On this web site, you can start your own participant brings one’s ego to the story, one’s
journey of exploration. Please read the own awareness, rather than identifying with,
Introduction and follow its suggestion for “How taking the part of, one or the other character in
Do I Start?” Feel free to contact me at: the story The more cross-cultural the story is—
[email protected] for example, all cultures are likely to have
This is an introduction to seminars designed creation stories, and stories about the origin of
for self-study. Each session or envelope contains science—the more universally valid or typical do
a story or direction for an exercise with the those characters seem, and the easier it is for the
material of the story. You, the participant in this hearer of the story to say, “Hey, that character IS
seminar, encounter the story or do the exercise, me, and that is MY story.” Try NOT to do that.
and then respond to one or more question(s) on The comparison of story with stage is quite
this web site. The response can be in writing or apt because as action on stage involves feelings
any other form of expression, and can be and emotions of onlookers, so encounter with
telephoned or sent by e-mail or post to: story can activate an individual participant’s
[email protected]. Rudy will inner knowledge and experience analogous to the
then send you the next session (e-mail) or story character(s)’ knowledge and experience.
envelope (paper) of the self-study seminar. The That can happen whether or not the participant
method is adapted for self-study from group had previously been aware of any feeling or
workshops using different stories and questions. actions corresponding to those of one or more
characters in the story. A shorthand phrase for
Method such activation is that one or another of those
If you were using this material at a group story characters is constellated in a participant by
workshop or seminar, you would be sitting in a the participant’s work with story and questions.
circle. Each person in the circle would hear from Another way of saying this is that no one
a facilitator what is on the web site as Session I, character in the story is or describes the whole of
or what is in envelope I in the paper version — a me, but it often describes a part of me. I may
short story, and then a question to which each not have been aware of that part of me prior to
member of the circle responds. It is not a my work with that story.

250
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Marcus, An Interactive Web Site for Ethics Training

Questions any time. Such changes are seen as fresh insights

If you encountered this material in a group rather than as error or shame for earlier choices.
workshop, the facilitator would ask one or more
questions after each story. Responses to the “I’ve heard that story before”
question are addressed not to the facilitator, and Of course you have heard that story before. It is
not to other members of the group but to the part of our cultural heritage. It is a myth. It is
center of the circle. The story is “myth,” a universal truth. You and I have that story in our
technical term that has been defined by Joseph bones.
Campbell as something that happened and But in the group workshop I am not listening
continues to happen to many people at many to, and in the self-study I am not reading, that
times and places, in contrast to history which story to find out how it ends, “who done it,” or
happened at one time and place to one person or what the facts are. I am called on to take a stand
a group of persons (and, a scientist would add, in vis-a-vis that story at this particular moment. I
contrast to science which is something that am listening to it, or reading it, to see whether I
happens and is replicable under specified still have the same answers to questions evoked
conditions). As a myth, the story is considered to by the story or to see whether the story evokes
bubble out of the center of the group circle rather new questions.
than from the facilitator. Similarly, the question Bruno Bettelheim has noted that a fairy tale
asked is considered to be coming from the center asked for most frequently by a child most likely
of the circle, and therefore responses are describes what that child feels to be its most vital
addressed to the center of the circle rather than to problem at that stage of its life. (For example,
a person at the perimeter of the circle or to the the Grimm Brothers’ fairy tale that gripped my
facilitator. That should minimize any onus on, or emotions for many decades was “The Goose
defensiveness by, a respondent about one’s Girl.” Like that protagonist, I also lost home,
response. Similarly, others in the circle are country, and native tongue as a pre-pubescent.)
expected to respect any response since it is Ritual has been described as a process of
addressed to the center and not to them. activating a myth. In that sense repeated
Responses are not to be discussed or argued with, exposure to a story—by rereading it, by noting
particularly not by the facilitator. whether responses to story and questions change,
The discussion circle consists of center and and by asking new questions about it—
perimeter. Without participation from the reconstellates powerful, often numinous,
perimeter, in the form of responses to questions, characters within myself.
the circle and therefore the seminar would not
exist. It is important for each participant to hear Instructions
oneself verbalize a response, even if that You are not at a group workshop now, so you
response sounds similar to one that has already have an opportunity to create your own pace and
been given. place. Find a private space and time for 45
The questions are designed to evoke choice minutes, turn off the telephone, put out the cat,
and feelings rather than to test knowledge or and open one of the stories. Read the story,
recall. Reasons for, or explanations of, choices consider it, and respond to the question(s) in
may be asked for. Respondents will be asked to writing or other art form. Send your response to:
stick to the subject matter of a question because [email protected]. Stay with
one of the easiest ways of escaping a choice or a the story for a day or more—preferably a week.
feeling is to talk about something else. Similarly, Look at the story, questions, and your responses
the question asked at a session is about the story occasionally, and write down any additional
told in that particular session, not about the end thoughts. Note any additional insights.
of the story or about another story. Each Your response will be acknowledged and the
question is a universe of discourse, embedded in next Session will be sent to you. Repeat the
the universe of discourse of the story of that process with the next Session. Continue in that
particular session. (By definition, a universe of manner until the final Session.
discourse is a separate world surrounded by an Because this kind of work is an ongoing
impermeable barrier.) A participant is free to process and new insights keep popping up, it is
state different choices, feelings, and opinions at well to keep the Session materials, your

251
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
responses, and the facilitator’s comments in a
notebook. You will find that collection a growing
resource as new insights arise. You will also find
that it becomes a valued friend and adviser in
dark times.

Which Story?
Four stories are available as Sequenced Self
Studies at this time. They are:
• Cracked Pot (11 sessions)
• Moses Mendelssohn’s Dream (5 sessions)
• Rainmaker (5 sessions)
• Becket (8 sessions)
Any of those Sequenced Self-Studies is worth
doing in its own right in the same way that one
goes to a movie or takes a trip for adventure,
enjoyment, or enrichment.
Just as movies or trips also may be taken
with specific purposes in mind, such as
information or education, these stories can be
used for specific purposes as well as in their own
right. For example, Cracked Pot and Moses
Mendelssohn’s Dream have been used for
working with self-worth problems. Rainmaker
has been useful for conflict resolution in highly
polarized situations. Becket is a good practicum
for finding moral advisers in hierarchically
structured organizations. Both Rainmaker and
Becket are excellent self-studies for ethics
training.

How Do I Start?
On the following pages [of the web site] you
will find the first Session of each of the available
self-studies. Choose one, follow the instructions,
and send your response to:
[email protected].
Rudy will then comment on your response
and activate the next session of your self-study.

252
III. Research Theory and Methods

8. Detection Methods
The Misuse of Statistics: Concepts, Tools, and a Research Agenda
John S. Gardenier, National Center for Health Statistics, Centers for Disease Control and
Prevention, USA

David B. Resnik, Brody School of Medicine, East Carolina University, USA

Key Words. Competence, Ethics, Journals, Research misconduct, Statistics

While it is widely recognized that the proper use of statistics is a key element of research integrity,
there has been considerable debate about how to understand or respond to the misuse of statistics in
research. To understand what is meant by “misusing statistics,” it is important to describe the role of
statistics in the scientific method and relate the concept of “misuse” to other ethical concepts, such as
“misconduct” or “incompetence” or “negligence.” We believe that some misuses of statistics can be
considered misconduct, although most misuses should be viewed as negligence or deficits of
competence.
Statistical methods, theory, techniques, and models play an important role in several stages of the
scientific method, but we will focus here on just two stages (See Figure 1). First, statistics is essential
to good experimental design as in randomized clinical trials, for example. In order to obtain a
rigorous test of a hypothesis, it important to obtain data that can provide evidence for or against the
hypothesis. If the hypothesis is a comparative or quantitative statement, such as “drug x is more
effective than drug y” or “less than five percent of patients suffer serious side effects from drug x,”
then the conclusions must be based on statistically significant results. For example, an experiment
that compares the effects of two drugs on only ten patients is very unlikely to produce statistically
significant results. If some or all of those patients are subjected to health risks in the experiment, this
creates two additional ethical problems. First, it is unethical to expose a human subject to an
unnecessary experimental risk, unless the potential benefits (to the individual or to society) of
exposure to the risk outweigh the potential harms. If the experiment is not well designed such that no
meaningful conclusions can be drawn, then the potential benefits will not outweigh the potential
harms. Second, when patients give informed consent to participate in research, they usually believe
that the research is valuable and may advance science. Encouraging or even allowing subjects to
participate in an experiment that is highly unlikely to yield valid results is implicitly deceptive. It is
important to address the statistical issues before conducting experiments or tests, because once one
has gathered and recorded data, it may be too late to correct statistical (or ethical) flaws in the design
of the experiment (1). The expression “garbage in, garbage out” applies here.
Second, statistics is important in an analyzing and interpreting data. There are many different
statistical tools that one may use to analyze data, ranging from simple procedures, such as t-tests and

Corresponding author: John S. Gardenier, CDC/NCHS, 6525 Belcrest Road, Hyattsville, MD 20782, 301-458-4229
(voice), 301-458-4021 (fax), [email protected].
* Disclaimer: This paper represents the personal views of the authors. It does not necessarily represent policies or positions
of their institutions. The research effort reported here was not supported by grant funds.
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Figure 1: The Role of Statistics in the Scientific Method
Define problems,
questions, and
research aims

Review the literature

Develop a hypothesis
Statistical methods,
measurement tools,
and models
Design experiments
or other tests

Collect and
record data

Revise or modify
protocol or hypothesis
Analyze and
interpret data

Disseminate results Peer review Replication of results

Public understanding Scientific impact

of research of research

linear regression, to more complex ones, such as in the data and the real signal, then someone who
analysis of covariance and statistical modeling. incorrectly uses statistics may produce a result
It is not our aim to discuss these methods here, that is distorted or even artificial. A person who
but we would like to point out that it is relatively correctly uses statistics will amplify and clarify
easy to misuse these methods. To apply any the signal without distorting it (2).
statistical method correctly, one must have With this understanding of the role of
information about the variables used (continuous statistics in research in mind, we can clarify what
or discrete, gaussian or bimodal, etc.), we mean by “misuse” of statistics. Not all
information about the sampling process used misuses have equivalent ethical implications, as
(sample size, independence, randomness, we discuss later. A “misuse,” for our purposes, is
representativeness, etc.), and a sound an incorrect use, i.e., a use of statistics that is not
understanding of the theory and assumptions appropriate, given the research question, the
underlying that method. If a researcher does not experimental design, and the methods being
use a method correctly, then conclusions may used. For example, it may be appropriate to
overestimate or underestimate an important exclude outliers if there is credible evidence that
relationship or effect. If we think of statistics as such points are not part of the statistical
a tool for distinguishing between random “noise” population represented by the sample. It may
256
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gardenier & Resnik, The Misuse of Statistics
also be appropriate to use statistical methods to requires adequate application of both statistical
fill in (or impute) missing data for the purposes and subject matter expertise to analyses. There
of statistical analysis. What’s the difference might be varying degrees of culpability in a
between appropriate and inappropriate exclusion failure to meet this criterion. Clearly, honest
of outliers or appropriate and inappropriate error is never misconduct. Neither is it
imputation of data? Many books on statistical misconduct when two or more well qualified
methods discuss these topics, but from an ethical statisticians or other researchers disagree about
viewpoint they boil down to the following: an technical issues in a given research protocol.
appropriate exclusion (or imputation) is one that Still, some misuses of statistics in research do fit
dampens the noise without altering the signal that the definition of misconduct used by the federal
describes the relationship or effect. government. That may be hard to establish by a
Misuses of statistics can also occur in the preponderance of the evidence. When a person
absence of erroneous or distorted results. Misuse changes or fabricates data, one at least has some
can also arise from a failure to provide the kind of record that one can use to imply intent.
research community with important information When a person manipulates analyses of data,
about the methods used or the experimental there may be no record to prove the manipulation
design. Researchers need to address such was deliberate or even culpably negligent. Thus,
statistical issues as excluding outliers, imputing as a purely practical matter, it may be very
data, editing data, “cleaning” data, or “mining difficult investigate or prosecute such cases (10).
data.”2 These practices are often practical, or
even necessary, but it is important to discuss The Importance of Correcting Statistical
them honestly and openly when reporting Misuse
research results (3). Statistics play vital roles in most aspects of
Thus, there are two types of misuses in modern post-industrial societies. Although
statistics: (1) using statistical methods, statistics are sometimes dismissed as trivia or
techniques, or models in ways that produce fuzzy math, distrusted as biased, or directly
distorted or artificial results; (2) failing to equated with lying, the truth is that they are
disclose important information about statistical inescapably important (11). As noted in the
methodology to researchers. Misuses of statistics Preamble to the Ethical Guidelines for Statistical
may (or may not) violate several ethical Practice:
obligations, such as the duty to be honest, the The professional performance of statistical
duty to be objective, the duty to avoid error, and analyses is essential to many aspects of society.
possibly the duty to be open (4). There has been The use of statistics in medical diagnoses and
considerable debate about whether “misuse of biomedical research may affect whether
statistical methods” should be classified as individuals live or die, whether their health is
misconduct (5). The federal government and the protected or jeopardized, and whether medical
scientific community have moved toward a science advances or gets sidetracked. Life,
death, and health, as well as efficiency, may be
narrow definition of misconduct that focuses on at stake in statistical analyses of occupational,
fabrication, falsification, and plagiarism (6, 7). environmental, or transportation safety. Early
The new federal policy implies that the misuse of detection and control of new or recurrent
statistics could be classified as a form of infectious diseases depend on sound
misconduct when it involves intentional epidemiological statistics. Mental and social
deception. Some misuses could be classified as health may be at stake in psychological and
“fabrication” if they involve making up data or sociological applications of statistical analysis.
results, or “falsification” if they involve Effective functioning of the economy
manipulating, changing, or omitting data or depends on the availability of reliable, timely,
results. Misuses of statistics that do not involve and properly interpreted economic data. The
intentional deception could be viewed as honest profitability of individual firms depends in part
on their quality control and their market
error, incompetence, bias, or “serious deviations” research, both of which should rely on
from acceptable practice (8). A person who statistical methods. Agricultural productivity
makes excessive errors due to haste, ignorance, benefits greatly from statistically sound
or sloppiness may be considered to be negligent applications to research and output reporting.
or lacking the needed degree of competence, Governmental policy decisions regarding
statistical or otherwise (9). Professionalism public health, criminal justice, social equity,
257
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
education, the environment, the siting of critical • Career ambitions or aspirations
facilities, and other matters depend in part on • Conflicts of interest and economic motives
sound statistics.
Scientific and engineering research in all
• Inadequate supervision, education, or training
disciplines requires the careful design and We believe that all of these factors probably play
analysis of experiments and observations. To a role in misuses of statistics, but our conclusions
the extent that uncertainty and measurement are merely speculative. More research is needed
error are involved – as they are in most research on this topic. However, we would like to discuss
– research design, data quality management, two other possible factors in the misuse of
analysis, and interpretation are all crucially statistics that are not on the above list of “usual
dependent on statistical concepts and methods. suspects.”
Even in theory, much of science and First, there are now many computer
engineering involves natural variability. programs that analyze data. These programs are
Variability, whether great or small, must be
very user-friendly; all you need to do is load your
carefully examined both for random error and
for possible researcher bias or wishful thinking. data set and choose your statistical test in order to
... get results. One may even run several different
Because society depends on sound tests in an attempt to increase the significance
statistical practice, all practitioners of statistics, level (or p-value), although this can invalidate
whatever their training and occupation, have the testing. While these programs save a great
social obligations to perform their work in a deal of time and effort, they may contribute to
professional, competent, and ethical manner. statistical misuse in that it is possible to plug
(12) some numbers into one of these programs
without knowing how the analysis works, or why
If researchers are careless or deceptive in their a certain test would (or would not) be an
use of statistics, harms and costs to society will appropriate test. We think this problem has a
result. Poor statistics in science leads to poor fairly obvious solution: teach more statistics in
science. The research record can be corrupted or research. If students and researchers understand
polluted, wasting the time and energy of other how to use statistics properly, then they should
researchers. At the very least, research funds lost have fewer problems using statistical computer
in bad research represent an opportunity cost in programs. Indeed, we believe that education is
that those funds could have been allocated to the key to improving statistical practice.
more deserving projects. Second, it has become standard practice in
For all of these reasons, it is important some areas of research to only publish results
that scientists and science administrators pay that have a p-value of 0.05 or less. The best
careful attention to the quality of statistics in journals use more comprehensive criteria
science as funded, performed, and reported in enforced by competent statistical peer review.
their areas of jurisdiction and of competence. We here address only those journals that place
Good statistical work should be defended when it excessive reliance on the p-value. The value of
is attacked inappropriately. Bad statistical work 0.05 is an arbitrarily chosen number; there is no
sound statistical or philosophical reason why a p-
should be detected and corrected as appropriate. value of 0.06 is fundamentally different from a p-
value of 0.05. However, under pressure to
What are the Contributing Factors to publish, researchers may decide to massage or
Misuse? manipulate data in order to obtain “significant”
There is not a great deal of evidence that has a results. Furthermore, there is now a growing
direct bearing on the misuse of statistics in body of literature on publication bias in research
research. However, if one assumes that many of (15-17). Publication bias occurs when there are
the factors that contribute to other ethical discrepancies between the published research
problems in research, such as misconduct, record and the complete research record. The
probably also play a role in the misuse of discrepancies occur because journals tend to
statistics, then one could cite the following publish only “significant” results. There are
factors, i.e., the “usual suspects” (13, 14). some good potential solutions to the p-value
• Pressures to publish, produce results, or problem. First, researchers should realize that p-
obtain grants values are merely conventional, not sacrosanct.

258
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gardenier & Resnik, The Misuse of Statistics
Second, they are also often sensitive to various research funding agency were to adopt a policy
theoretical assumptions and may give erroneous of stating in grant announcements that all grant
results due to mere artifacts of a data sample. proposals received for projects employing
Third, not all statistical computer packages statistical methods would be expected to be
compute all tests correctly. Fourth, journals performed in accordance with the Ethical
should be willing to publish results that are Guidelines for Statistical Practice, that would put
substantial contributions to the literature of the real moral pressure on both proposers and
field, not just those that appear to have met a grantees to avoid misuse of statistics. If journal
conventional statistical test. The test result editors were to state in notices to authors that any
reported may not be correct, and even a correct papers containing statistical methods submitted
conclusion that a certain hypothesis was not to that journal would be implicitly subject to
statistically supported by data from a well- those same guidelines, some of the authors would
designed study may be useful in limiting future be more highly motivated to avoid misuse of
fruitless research by others. Finally, researchers statistics.
and research organizations should create If all scientists and engineers who are
databases for unpublished data of archival value competent in statistical methods would note
and make those data publicly available (18). published examples of misuse of statistics and
report those to the funding agencies or journal
Statistical Ethics, a Powerful Tool for editors involved, then the recipients would
Research Integrity become more motivated to enforce sound
Statistical ethics is a relatively recent statistical practice. In short, we should not let
development. The seminal work, by W. Edwards ethics documents sit unused on shelves or in
Deming, was first published in 1965 (19). The unvisited cyberspace. Ethical considerations
American Statistical Association developed a have practical consequences for good or evil.
series of statistical ethics codes or guidelines The failure of good people to use them
starting in 1979. Their current official Ethical effectively contributes to the likelihood that other
Guidelines for Statistical Practice was people may perpetuate statistical misuse either
promulgated in 1999 (12). The International through intent to deceive or simply through
Statistical Institute instituted its Declaration on deficits of statistical competence.
Professional Ethics in 1985 (20). The United
Nations has published Fundamental Principles of A Proposed Research Agenda
Official Statistics in the early 1990s, the current While we believe that there are still many
official version being dated 1994 (21). important conceptual and theoretical issues
The pattern that emerges from this brief relating to the use/misuse of statistics in research,
history is that initial efforts to approach the issue it should be clear from this brief discussion that
tend to be optimistically simple. Corrections more empirical research is required on the
over time add to the scope and complexity of the incidence of statistical misuse, its causes and
documents. The most recent document breaks effects¸ and on the efficacy of using ethics
out areas of ethical responsibility for all people education and ethics documents as tools for
using statistical methods professionally (12). It improvement. The following are some of the
covers separately, for example, responsibilities in empirical research questions we think are
publications and testimony, responsibilities to important to study:
funders or employers, to research subjects, to 1. How many (or what percentage of) pub-
research team colleagues, and responsibilities lished studies make statistical mistakes?
regarding allegations of misconduct. Beyond 2. How many allegations of research miscon-
addressing responsibilities of the individuals, duct involve misuses of statistics?
moreover; it also addresses the responsibility of 3. How many researchers believe that the
those employing practitioners of statistical
misuse of statistics is an important ethical
methods to provide a suitable moral climate for
that work. issue in research?
Such statistical ethics documents become 4. Do different fields have different statistical
tools for research integrity when they are integral practices or take different approaches to the
to actual practice. For example, if a federal misuse of statistics?

259
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

5. What is the incidence of publication bias in 7. Office of Science and Technology Policy. Federal policy
various fields? on research misconduct. http://www.ostp.gov/html/
001207_3.html , 2000.
6. What do researchers and students know 8. Panel on Scientific Responsibility and the Conduct of
about statistics? Research. Responsible Science: Ensuring the Integrity
7. Where, when, and how do students learn of the Research Process. Washington: National
about misuses of statistics in research or Academy Press, 1992.
9. Grinnell F. The Scientific Attitude. New York: The
other ethical issues in statistics?
Guilford Press, 1992.
8. How often do researchers use statisticians 10. Dahlberg, J. Personal communication. Office of
or other statistical consultants? Research Integrity, Rockville, MD. 20 November 2000.
9. Are editors and reviewers able to catch 11. Huff D. How to Lie with Statistics. New York: WW
statistical misuses? Norton, 1954.
12. American Statistical Association. Ethical Guidelines for
10. Can data audits detect misuses of statis- Statistical Practice. Alexandria, VA: American
tics? Statistical Association, 1999. Available at http://
11. Do research ethics codes or policies amstat.org/profession/ethicalstatistics.html.
address misuses of statistics? 13. Macrina, F. Scientific Integrity. Washington: American
Society for Microbiology Press, 2000.
12. When ethics education or ethics documents
14. Commission on Research Integrity. Integrity and
are used as tools to improve research Misconduct in Research. Washington: Public Health
integrity, how effective are they at promot- Service, 1995.
ing the proper use of statistics? 15. Ioannidis J. Effect of the statistical significance of
13. How often do institutional review boards results on the time to completion and publication of
randomized efficacy trials. JAMA 1998; 79: 281-86.
(IRBs) discuss statistical issues in human 16. Stern J and Simes R. 1997. Publication bias: evidence
subjects research? Do IRBs use statisti- of delayed publication in a cohort study of clinical
cians? research projects. BMJ 1997; 315: 640-45.
14. How do misuses of statistics affect the 17. Easterbrook P, et al. Publication bias in clinical
research. Lancet 1991; 337: 867-72
public? Do such misuses ever cause harm
18. Rennie D. Fair conduct and fair reporting of clinical
to the public or threaten public health or trials. JAMA 1999; 282: 1766-68.
safety? 19. W. Edwards Deming. Principles of Professional
15. How often do statistical issues arise in Statistical Practice. The Annals of Mathematical
public policy debates? Statistics 1965; 36: 1883-1900.
20. International Statistical Institute. Declaration on
16. What does the public know (or not know) Professional Ethics. 1985. Available at http://
about statistics? www.cbs.nl/isi/ethics.htm.
17. How do lay people interpret important 21. United Nations. Fundamental Principles of Official
statistical concepts, such as “probability,” Statistics. 1994. Available at http://www.cbs.nl/isi/
fundamental.htm .
and “risk”?

Bibliography
1. Johnson R. and Bhattacharyya G. Statistics: Principles
and Methods. New York: John Wiley and Sons, 1985.
2. DeMets D. Statistics and ethics in medical research.
Science and Engineering Ethics 1999; 5: 97-117.
3. Resnik D. Statistics, ethics, and research: an agenda for
education and reform. Accountability in Research 2000;
8: 163-88.
4. Bailar J. Science, statistics, deception. Annals of
Internal Medicine 1986; 104: 259-60.
5. Guston D. Changing explanatory frameworks in the
U.S. government’s Attempt to define research
misconduct. Science and Engineering Ethics 1999; 5:
137-54.
6. Office of Research Integrity. Federal policy on research
misconduct. Federal Register: October 14, 1999; 64,
198: 55722-55725.

260
Images as Evidence: Forensic Examination of Scientific Images1
John W. Krueger, Division of Investigative Oversight, Office of Research Integrity, DHHS,
USA

Keywords: Autoradiograms, Blots, Image processing, Manipulation and falsification, Scientific misconduct

A “questioned” scientific image, i.e., suspicions of falsification (or plagiarism) of image data, such as
photographs of PAGE gels, autoradiograms, and blots (Western, Northern, and Southern) can give rise
to an allegation of misconduct in science. Pursuing oversight review of institutional investigations
and reviewing allegations that ORI receives directly, ORI commonly examines the evidence through
image processing. Typically, the examination can extend beyond merely asking “what is the evidence
the image is/isn’t authentic?” and/or “are two contested images really the same?” Examples from
these cases illustrate the general principles in forensic image processing and several methods that ORI
has found useful in resolving the questions at hand. They provide an opportunity for further
instruction as to what constitutes data falsification in an image.

Design/Methods
Source of Material: The material for this presentation was taken from a survey of 19 ORI cases
that involved allegations of falsification or plagiarism of the images of gels, blots, auto-radiograms,
and micrographs. The cases span a period from 1990 to 2000. The number of such questioned image
allegations has generally increased, as has their incidence relative to other ORI cases. (Figure 9) A
compilation from this review is discussed below.
Software: Most of ORI’s image analysis was done on a Macintosh® computer. The reason is
both historical and practical; files transfer easily from the Windows® platform to the Macintosh®;
but the opposite is not always true.
ORI has found several different image processing programs that are readily available and well
documented so that the results can be easily shared with all parties in a potentially adversarial dispute.
(1, 2) Each separately —or in combination with the others— offers distinct advantages. The image
processing was conducted using either NIH Image (3) and/or Adobe Photoshop® (4), both of which
were equipped with the Image Processing Tool Kit® (IPTK) plugins. (5) NIH Image, developed at
the National Institutes of Health, is in the public domain and is ideal for analytical treatment of 8 bit
(256 shades) monochromatic images. Photoshop is better suited for conducting overlay comparisons
of two images and for working with color, but it requires the IPTK’s plugins for analytical work.
Finally, ImageJ (6) is an update of the NIH public domain software that is compatible across
computer platforms and will process images at 8, 16, and 32 bit depth; thus, it can detect vastly fainter
features that might be hidden in the image.
Other Resources: Articles that can serve as guidance to issues involved in the forensic
examination of contested documents can be obtained on the Internet. (1, 2) Those sites can serve as

Corresponding author: John Krueger, Office of Research Integrity, 5515 Security Lane, Suite 700, Rockville, MD, 20852,
301-443-2263 (voice), 301-594-0039 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
produced, such as the sources, dates, and
incidence of re-use, may establish whether a
pattern of misrepresentation existed that rules out
honest error. Examples from ORI’s cases
illustrate these points.
Figure 1 represents a photographic mock-up
of Western blot data, consisting of five
photographic strips, in which the 2nd to 4th lanes
were on one strip. Although purportedly
showing differentdeterminations of protein
created by separate mutant gene constructs, the
1st, 4th, and 5th lanes look unexpectedly similar,
but it is difficult to say for certain that they are
the same.
One generic principle in making comparisons
to determine the authenticity of data is to look at
the features that would otherwise be un-
noteworthy, such as fine features hidden in the
background.3 There may be random features that
are hidden from our perception. The human eye,
which responds to contrast, can distinguish only
~50 shades of gray (7) or less (8), but it can
detect 100 shades of color (8).4 However, the
computer's response is not dependant on contrast;
Figure 1. Original Western blot data. The results of an it can selectively amplify very slight differences
electrophoretic mobility shift assay to show bands reflecting
in shade. The ability to detect such differences
the gene expression of separate mutant proteins. However,
the shape of the bands and the pattern of the background in
can be affected by the “depth” used to digitize
the 1st, 4th, and 5th lanes look alike. the image, which in this case is 256 shades of
gray.5 The amplified differences in gray shades
links to find other material. can next be shadowed and assigned false-colors
to make faint differences even more visible, as
Reasons for Examination and Some shown in Figure 2.
Principles of the Image Analysis Methods These steps reveal faint artifactual features
The usual motivation for image analysis is to that were “hidden” in the background which are
examine the authenticity of a particular document common to three of the lanes. Thus the
or to determine whether two purportedly respondent's claim, that at least two of the three
different images really were derived from the lanes (1, 4, or 5 in Figure 1) represented evidence
same experiment.2 In fact, image analysis for gene expression of different mutant proteins,
provides information that addresses other issues. was a clear falsification of data.
For example, features can be detected that reveal Enhancement of the small difference in
the source of the image, whether it is compatible shades can also expose minute structural details
with laboratory records such as autoradiograms in the morphology of bands, which otherwise
or prior blots (see note 2), and whether the would look smooth and featureless. Figure 3
questioned image existed on a computer as a file, illustrates a photo-montage from the above case;
or on a website as a component of someone the bands appear similar in the 1st and 5th lanes.
else’s homepage. Second, the analysis of the Contrast enhancement and false-coloring of
latter sources can provide dates of creation, the above image as shown in Figure 4
which can be corroborated with laboratory demonstrate that the respective bands share
records, etc. Third, image enhancement may similar miniature features. Thus, the image
reveal evidence for the mechanics of the figure’s analysis showed that the first and the last lanes
construction, such as edges of photographic were from the same experiment.
prints and presence of “white-out” and may In both examples above, the falsification was
uncover “hidden” details, such as erasures of associated with false labeling of data that had
labels. Fourth, an analysis of the new facts been “re-used” from another experiment. The
262
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Krueger, Images as Evidence
Figure 2 (left). Image enhancement of the questioned
Western blot data. This ORI image analysis figure shows
only the 1st, 4th, and 5th lanes from Figure 1. Contrast
enhancement of the monochromatic gray-scale image,
followed by shadowing and false-coloring (using NIH
Image), revealed small features in the background artifact
that are common to all three lanes (arrows) which the
respondent had falsely represented as different. Note that
in this case some differences can also be found, such as an
horizontal artifact under the top band in the 4th lane, but
they are in the background and represent artifacts that were
introduced at some later point.

Figure 3. Western blot data. The results purportedly found

a good agreement between the observed and the predicted
size of five mutant proteins. However, the 1st and the 5th
lanes’ bands look similar.

Figure 4. Image enhancement of the 67 kDa MW and 32 kDa MW bands from Figure 3. The bold lines denote
miniature features in the bands’ morphology that indicate both were actually the same data, which the respondent had
falsified in re-use.
263
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
second example showed an additional
falsification involving a false claim that the
molecular weights had been determined. In this
case, the intent to falsify the data is prima facie,
because the molecular weight could not have
been measured for the last re-used band. Finally,
because the molecular weights were purported to
approach the predicted values, the evidence also
indicates that the falsifications are significant.
These elements strengthen the findings.
Background detail and miniature features
cannot be examined by image enhancement in all
specimens. Fortunately, numerous other
approaches are available in image processing to
compare two questioned images. In general a
combination of methods is determinative. For
example, the morphology, size, and vertical
arrangement of the bands and the existence of
larger artifacts are the most obvious features to
compare. Moreover, the horizontal spacing
between the bands should not be overlooked;
because substances rarely migrate on gels
absolutely parallel, there may be slight Figure 5. Overlay of the mentor’s Northern blot mRNA
differences in lateral disposition that are also data (small rectangle) with a figure from a different
significant. Some forms of artifact might re- experiment from the student’s thesis (tall rectangle). In this
occur, such as that introduced by a faulty film ORI image analysis, the actual fit was determined
dryer and/or the edge of a blot on an mathematically and showed the missing blot actually had at
autoradiographic film. The key question in cases least seven lanes, indicating the respondent’s claim was
false.
of “replicating” artifacts is whether a relationship
to other features should exist. of six lanes, separate by an empty lane.
How to best visually represent the results of However, the overlay, shown in Figure 5, which
an image overlay is always a challenge. A was established as the best mathematical fit
visually effective and efficient method is to between the two sources, demonstrated that the
overlap color-coded plots of the “contour” map missing original blot had to have had at least
of the intensities in two separate blots, where the seven lanes. Thus, the proffered laboratory
areas of overlap generate a third color. If two records could not be evidence of the mentor’s
gray scale images are overlaid, the interpretation “missing” data.
of the origin of features in the overlay becomes Analysis of Poor Images: The poor quality of
problematic unless each is first converted to a an image is not necessarily a deterrent to the
suitably chosen monochrome color scheme. application of the above tools to its examination.
Reconstruction of a Missing Document: The left side of Figure 6 shows a poor quality
Analysis of an image can also be used to test the photocopy of data that was submitted in a
proffered source of a questioned image under mentor’s National Institutes of Health (NIH)
circumstances in which the original raw data are grant application, which purported to be a
missing. Figure 5 represents a composite image, Western blot of an immunologic protein, “P-48,”
which was created by combining a figure of a using 125I-labeled human lymphocytes. The
questioned Northern blot in the mentor’s figure on the right side of Figure 6 represents the
manuscript with a figure of a different enhanced image of an autoradiogram from his
experiment shown in the student’s thesis. student’s experiments, which used 35S -
Unfortunately, the original blot and its methionine labeling of cultured rats cells.
PhosphoImager computer file were missing, but The distracting artifact due to over-
the mentor provided laboratory data purporting to photocopying could be minimized by image
be a different representation of the same blot (an processing. This step revealed additional bands
ethidium bromide stain) that showed two groups in the former with more certainty, and it more
264
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Krueger, Images as Evidence

Figure 6. Examination of a poor quality photocopy. The mentor submitted the left hand “125I-labeled” figure in an NIH
application. At right is shown the student’s 35S-labeled autoradiogram, in which the band pattern was shadow-enhanced
(green arrows). An artifactual lane distortion is denoted by the red arrows, which is weakly indicated in the photocopy.

clearly exposed a similar artifactual distortion of falsified the preparation, the experimental
the lanes, as shown in Figure 7. The mentor had conditions, and the molecular weights in the
photocopy that he had submitted to the NIH.
Recovery of Probative Details:
Examinations of images may even reveal new
evidence that bears upon other elements that are
required for a finding of scientific misconduct.
In another case, the allegation involved six
instances where different sets of autoradiograms
were allegedly falsely labeled and presented as
different experiments. The student claimed these
were honest errors, due, in part, to her
inadvertent use of unlabeled autoradiograms.
However, image enhancement by one of the
institution’s committee found evidence that the
original label on one autoradiogram had been

Figure 7. Computer enhancement of the bad photocopy

shown in Figure 6. In ORI’s image analysis, the distracting
artifact in the photocopy can be removed by filtering, while
false-coloring further enhanced the bands. The lane
distortion artifact, present in the student’s autoradiogram
(Figure 6) was apparent in the same relation to the bands in
the enhanced image, showing the student’s results were
falsified by the mentor to NIH.
265
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Figure 8. An example from one of six sets in which an autoradiogram had been falsely labeled and re-used. The
institution’s image analysis found evidence that the label for the prior experiment had been erased on the corner of the
autoradiogram. The visible ink is blue, while evidence for the enhanced erasures is shown in red. Originally barely visible
only as a faint and diffuse blue smear, the erased label differed from the film’s background by only one gray level out of 256.
The erasures were visualized here by ORI, after the film had been scanned at 42 bit resolution and the erasures had been
selected for enhancement using their hue. The erased “MyoG” and “MyoD” denoted experiments on chickens and not
mice. Thus, the re-use of the autoradiogram could not have been an honest error from mixup of unlabeled films, as the
respondent originally claimed.

erased, but not fully (Figure 8). Thus, image subsequent analysis of figures in publications
processing revealed evidence that the found that there was a pattern as to the six
falsification was not an honest error. ORI’s instances of re-use that was not consistent with
their selection by chance.
A scientific image is
simply a picture constituting
evidence that a test was
carried out and/or that the test
produced a certain outcome.
In this context, the image is
construed as qualitative
“data.” It could also be the
basis for quantitative
measurements, i.e., by
measuring the “size” of a
substance, or as the raw data
for determine the amount of a
substance. Thus, one
consequence of discovering
the falsification in an image is
that there may be false claims
elsewhere in a paper.
Figure 9. Incidence of 19 ORI cases involving contested scientific images. The
data reflect when ORI’s case file was opened; this formal step can occur at any Compilation of
phase in a case’s history (i.e., at the allegation assessment, inquiry, or investigation Information from 19 ORI
stages). Thus the act of misconduct differs slightly in its date of occurrence. The
image analysis cases
percentages indicate the fraction of all ORI cases opened in those years.
“Tamper” refers to allegations where the intensity of bands was selectively
In all of the cases above, the
altered. “High-Tech” indicates manipulation by a computer to modify the image. questioned image qualified as
266
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Krueger, Images as Evidence

Type of Misconduct Number of ORI Cases Image Source Respondent

Alleged Thesis (student) 8
Falsely label as a 13 (5 students)
different experiment (3 mentors)
(re-use) Others (plagiarized) 3
Falsify molecular weight >13 Prior publication (self) 2
marker positions Status:
Cut, Graft, and Reuse, 5 Senior Faculty 7
alter lane positions to Junior Faculty 4
fabricate data Fellows 3
Tampering: selective 4 Students 5
manipulation of Allegation Source:
image content, Student/Mentor/Co-Invest. 9
enhance/subtract Reviewers 5
bands, move position Inquiry Committee 2
Plagiarism of images 3 Audiovisual Technician 1
(from Internet or Audience 1
journals), with false
Table 2. Characteristic of allegations of falsification of
claims of element(s)
images in 19 ORI cases.
from above
figures copied from published journal figures or
Table 1. Falsification of data with images–compilation from by use (and falsification) of images obtained
review of 19 ORI cases. This compilation indicates the from the Internet homepages of other scientists.
incidence as number of cases, which under- represents the Other aspects of these image-related
instances of misconduct, i.e., the actual number of figures or allegations are described in Table 2. Thesis
publications involved. The impact of the acts in each case
was, in some cases, dramatically higher; one case involved
research appears to provide a relatively frequent
40 separate figures and multiple publications. source of questioned images, falsified by both
students and mentors. In three cases, the images
data, the intentional fabrication or falsification of were allegedly obtained from others, and in two
which is the key element of research misconduct. other cases they involved falsification of images
On three occasions, a component of the that had been published earlier by the same
allegation also involved plagiarism. The laboratory. The source of most of these
allegations reviewed by ORI generally involved allegations was co-workers, although in five
use of fairly low-tech mechanisms for cases it was a reviewer who recognized the
misrepresenting the data (Figure 9), such as re- image as being from another source, or saw
use with false labels; in one case there were intrinsic evidence that the image could not be
multiple instances of use of a felt-tip pen to add a authentic. Most allegations did not arise because
band. Use of a computer to alter the content of the images looked inauthentic, but simply
the image has been seen less frequently.6 because they were either recognized as false or
Table 1 compiles the nature of the represented claims that a reviewer frankly
misrepresentations involving questioned images disbelieved. The questioned image was often
in 19 ORI cases. The most common allegation the one concern in a case that could not be easily
was falsification of data by misrepresenting the dismissed.
results as being a different experiment, which
also includes the attendant falsification of Discussion
molecular weights. Only five examples occurred The facts uncovered by the forensic examination
in which the lane markers were cut, re-grafted, of questioned images can often be corroborated
and shifted so as to fabricate a test that was never by other evidence, such as absence of records/
run. Purposeful tampering with the image to experiments on the requisite date(s), the
selectively enhance or remove bands has existence of dated computer files or other
occurred, but it was not very common. The versions, parallel images in publications, etc. In
allegations of plagiarism involved falsification of addition to the basic image processing, a clear
267
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
follow-up analysis is important. Bibliography
The most useful analysis of questioned 1. Blitzer, Herbert, “Digital Imaging,” Institute of Forensic
scientific images is done with a clear Image Processing, Purdue School of Engineering and
understanding of the experiment in question. Technology, IUPUI. PDF document available on the
This often requires collaboration between the Internet at http://www.engr.iupui.edu/ifi/articles.
Published in Law Enforcement Technology, March
image analysis and individuals who have a direct
2000.
familiarity with the conduct of the scientific 2. Scientific Working Group for Imaging Technologies,
experiments at issue. (9) To date, only two “Definitions and guidelines for the use of Imaging
institutions have reported to ORI using a Technologies in the Criminal Justice System” (vs. 2.1,
computer-based image analysis. Only one June 1999). Department of Justice PDF document
institution documented those results; in that available on the Internet at http://www.for-swg.org/
instance, image processing by a committee swgitin.htm.
member uncovered details that were 3. NIH Image is public domain software that was
determinative (see Figure 8). The information developed at the National Institutes of Health. A large
number of analytical macros for NIH Image are
from ORI’s cases indicates that most allegations
available via an active electronic bulletin board. NIH
involved “reuse” of the image to represent data Image is particularly useful for any quantitative and
from a purportedly different experiment. analytical measurements, false-coloring to better
Occasionally, photographs of gels or blots were visualize or code details of processed images, but it is
“cut and pasted” together in different limited to 8 bit gray scale images on the Macintosh.
combinations. Manipulations by computer were NIH Image is available on the Internet at http://
less common. rsb.info.nih.gov/nih-image/.
An image by itself creates a mantle of 4. Photoshop is a popular commercial program for both
authenticity, if only because we give unusual computer platforms (Windows and Macintosh).
Photoshop records the history of the computer’s actions,
weight to what we see. Yet in those cases where
to document the steps used in the image processing, but
scientific misconduct was found, discovery of its analytical capabilities are limited. Photoshop can be
one falsified image often led to the discovery of obtained from Adobe Systems, Inc., San Jose, CA
another, and in all the “original’ laboratory 95110-2704.
records were “missing.” Thus good laboratory 5. Photoshop-compatible “plugins” are available
practice may help to deter or to minimize the commercially as the Image Processing Tool Kit®
impact of falsification. (IPTK), Reindeer Games, Inc., 20 Battery Park Ave.,
Suite 502, Asheville, NC 28801. IPTK is also
Notes compatible with NIH Image; it greatly extends the
1 analytical capabilities of either program for analysis of 8
Any views expressed in this article are my own and do not
bit gray or 24 bit color images. Information can be
necessarily reflect those of the Office of Research
found on the Internet at http://members.aol.com/
Integrity. The citation of items in this article does not
ImagProcTK.
connote a product endorsement.
2 6. The base in support for analytical macros for ImageJ is
The questions are not limited to examining items that
not as large as the forerunner, NIH Image, but it is
look alike. For example, immunoblots from the same
growing. ImageJ and its plugins can be obtained on the
gel can be stripped and re-probed with entirely new
Internet at http://rsb.info.nih.gov/ij/plugins.
labeled antibody to reveal different protein bands.
3 7. Inoué, Shinya, “Video Microscopy,” Plenum Press, New
The forensic value of the background information is
York, N.Y. 1986. pp 80-83.
completely analogous to the basis for numerical forensic
8. Russ, John C. “The Image Processing Handbook,” CRC
analyses developed by Dr. James Mosimann in another
Press, Inc., Boca Raton, 2nd Edition. 1994. p. 213.
presentation at this meeting.
4 9. Indeed, the examples presented in this paper represent
A simple “thought” experiment makes the point more
the collaboration and assistance from my ORI
elegantly than examining the underlying physiology of
colleagues, Drs. John Dahlberg, Kay Fields, and Nancy
visual perception: any two gray levels, so alike that they
Davidian.
could be fairly represented as one shade, could still be
assigned two separate colors, say red and blue, of the
same intensity. (8)
5
Notice that digitizing at greater bit depth, such as 12 bit,
would in principle detect fainter differences in shading
to 1/4096 parts, rather than the 1/256 parts shown here.
6
It is debatable as to whether it would be more or less
difficult to detect computer alterations. What can be
said is that an allegation rarely arose because an image
on its face appeared inauthentic.
268
Terminal Digits and the Examination of Questioned Data
James E. Mosimann, Office of Research Integrity, DHHS, USA
John E. Dahlberg, Office of Research Integrity, DHHS, USA
Nancy M. Davidian, Office of Research Integrity, DHHS, USA
John W. Krueger, Office of Research Integrity, DHHS, USA

Keywords. Digits, Research misconduct, Statistical forensics, Terminal Digits, Uniform distribution

Our objective is to illustrate the use of statistical methods to examine the authenticity of data in the
investigation of research misconduct. We present examples of statistical analyses of questioned data
from several cases that illustrate the experience of the Office of Research Integrity. We show that the
statistical examination of numbers that are normally unrepeatable when experiments are repeated, or
otherwise are of inconsequential meaning, may reveal substantial clues as to the authenticity of
questioned data when compared with numbers in data that are unquestioned. We illustrate the
occurrence of the uniform distribution of non-leading (insignificant rightmost) digits in unquestioned
numbers, along with examples of deviation from such uniformity for fabricated or falsified numbers.
(Most people are unable to choose digits randomly.) We describe several cases in which a variety of
anomalies in data sets provided the impetus for the examination of rightmost digits. The anomalous
behavior of rightmost digits, when added to testimony and other physical evidence, can greatly
enhance or decrease the credibility of witnesses. The cases discussed involve: 1 and 2, Anomalous
behavior of terminal digits in published or recorded numbers; 3, Terminal odd digits in event times
that should have exhibited only even digits (and why); and 4, Data that were falsified by calculations
from computer spreadsheets (detected by the inclusion of an additional digit of accuracy).

Introduction
Allegations of research misconduct1 often are of the form that a particular experiment was not done
as described, or not done at all. In considering such allegations it is often necessary to examine
“questioned” data. Such data can establish that the experiment was performed as described.
However, if the allegation is true, then these questioned data are necessarily falsified or fabricated.
A useful way to assess questioned data is to examine inconsequential components of data sets that
are not directly related to the scientific conclusions of the purported experiment. Thus if the
allegation is true and the data are falsified, the falsifier typically devotes attention to numbers that
establish the desired scientific outcome. Properties of the numbers that are not directly related to the
desired outcome are less likely to receive consideration by the falsifier.
The same principle of examining details inconsequential to the scientific outcome appears valid

Corresponding author: J.E. Mosimann (present address), ABL Associates Inc., 5 Balmoral Court, Rockville, MD 20850,
[email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
whether the data are expressed in non-numeric Case 1: Uniformly distributed rightmost
form (images, written descriptions) or as digits in scintillation counts
numbers. Here we consider several cases where In the first case, experimental measurements
the data are numeric and lend themselves to were known not to have been done because
immediate statistical description. radioactive spots on the experimental sheets had
In all these cases we stress the importance of not been excised and hence could not have been
comparing “questioned” data with similar counted in the scintillation counter. Yet the
unquestioned data from the same laboratory or respondent’s notebook contained (falsified)
individuals. handwritten counts for that experiment. In this
case, faced with the evidence, the respondent
Rightmost digits admitted to the falsification of the numbers in the
Consider counts of radioactivity for a biological notebook.
preparation; for example, 5071. In a recount of In addition to the questioned counts, the
the sample, or in a replication of the assay, it is notebook contained handwritten counts that were
highly unlikely that the rightmost digits will be supported by counter output, and thus not
the same. Thus with two repetitions of the falsified. Both questioned and unquestioned
experimental procedure, instead of 5071, one numbers occur in pairs (a numerator and
might obtain respectively, 5109 and 4966. The denominator) and have large numbers of digits
rightmost, non-leading digits of these three (Table 1).
numbers are not the same. Thus _071 differs The following procedure was used to find
from _109, and in turn both differ from _966. digits. The rightmost digit of a number was
Digits are often recorded well beyond the designated as occupying “Place 1,” then the digit
repeatability of the experimental procedure. For to its left occupied “Place 2,” etc. Digits were
such rightmost digits, theoretically2 there is a examined in four places for each number, except
tendency to be uniformly distributed as expected that the leftmost digit was never included in the
in a lottery. For example, a uniform distribution analysis. Thus by way of example, the
of digits is expected in the Maryland Lottery. underlined digits would be included in the
Figure 1 shows the frequencies of the digits 0 to analysis: 1078, 251183, 735, 62034. It is clear
9 found in 5,106 winning “Pick-3” numbers (of 3 that a three-digit number contributes two digits
digits each) for the past ten years.3 This for analysis and a four-digit number, three digits.
distribution is not significantly different from Numbers of five or more digits contribute four
uniform. All digits have occurred with nearly the
digits.
same frequency, as they should in a lottery.
Chi-Square tests for uniformity of digit
distributions from 252 falsified counts from
notebook pages 141-152 are presented in Table 2.
The distributions are not uniform. Three of the
four Chi-Square values have probabilities less
than .05, and when digits from all four places are
grouped together, the total distribution is far from
uniform (Chi-Square = 30.94, df = 9, p=.0003).
Chi-Square tests for uniformity of the digit
distributions from 222 unquestioned counts also
are presented in Table 2. The distributions are
not significantly different from uniform. All of
the four Chi-Square values have probabilities
greater than .05, and when digits from all four
places are grouped together, the total distribution
is not significantly different from uniform (Chi-
Square = 11.09, df = 9, p=.27).
The unquestioned counts have uniform or
nearly uniform rightmost digits, whereas the
Figure 1. Ten years of Maryland Lottery Pick Three Digits, falsified counts do not.5
January 2, 1990 to December 31; 15,318 digits.

270
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data

Falsified counts Unquestioned counts

(Notebook page 145) supported by counter
printouts
(Notebook page 135)
Numerator Denominator Numerator Denominator
1078 251183 82267 170679
1770 217763 105584 190994
1091 225853 87592 181133
1434 238995 83341 197822
1247 241139 88426 172062
1131 260074 105068 194570
54350 220938 90707 150614
Table 1. Illustrative falsified and unquestioned counts from the respondent’s laboratory notebook.
Numerator (summation of reaction produced counts) and denominator (residual substrate count)
are associated with a given clone, and activity is expressed by the ratio, numerator divided by
denominator. Note that the 28 counts illustrated each contain from four to six digits.4

Chi-Square Results
For Falsified and Unquestioned Counts
Digits from 252 Falsified Counts
Place 4 Place 3 Place 2 Place 1 Total
Number 185 250 252 252 939
Chi-Square 34.8 29.3 13.2 27.1 30.94
D. Freedom 9 9 9 9 9
Probability .00006 .00058 .1521 .0013 .0003

Digits from 222 Unquestioned Counts

Place 4 Place 3 Place 2 Place 1 Total
Number 195 218 222 222 857
Chi-Square 14.3 9.89 8.72 11.33 11.09
D. Freedom 9 9 9 9 9
Probability .11 .36 .46 .25 .270

Table 2. Chi-square results for tests of uniformity of digit frequencies for falsified and unquestioned
counts. The rightmost place is “Place 1”; the next place to left is “Place 2” etc. (Leftmost digits of
numbers were excluded, so there are fewer “Place 4” digits than “Place 3,” etc.)

Case 2: Unlikely Patterns in Rightmost rows in each half represent, respectively,

Digits. different bacterial sources for the endotoxin.
In this case, we again demonstrate the ability of LPS was added at various concentrations to the
uniformly distributed digits to distinguish cell cultures. Thus the five columns of the Table
questioned from unquestioned data. However, represent different levels of LPS (left to right,
the digit analyses lead further to the identification respectively: 5000, 500, 50, 5, and .5 ng/ml).
of unlikely patterns in numbers that should not be The upper half of Table 3 represents cultures to
related, given the purported experiment. which endotoxin and stimulator cells were added
Table 3 (next page) reproduces the means at the same time. The lower half represents
and standard deviations from a published Table cultures to which endotoxin was added 24 hours
that was challenged by a coworker. prior to the addition of stimulator cells.
Lipopolysaccharide extracts (LPS) were purified However, while supporting notebook data could
from endotoxin from various bacteria. The five be found for the first four columns, no supporting
271
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Column 1 Column 2 Column 3 Column 4 Column 5

Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev
17697 1739 17399 1680 15085 1342 18262 2934 27191 1404
20164 3540 16746 1171 19397 1133 17889 3919 26999 7107
23323 3861 24154 722 19094 1340 28763 3373 28611 967
24474 4042 18918 4459 14224 828 24596 6327 29152 1407
29711 1519 21855 8458 23840 1695 29669 3222 28765 7104

24752 1455 22498 4591 21639 1347 32825 3063 70714 2106
32683 8535 26321 2753 20015 2020 34030 3917 68177 7155
43411 4682 41980 1705 34026 3906 47703 1894 66004 3924
26535 2349 41592 5699 31262 2796 54588 5065 74316 2192
33216 3762 37036 2071 27513 5062 32033 8307 71117 6817

Table 3. Published Table (Column 5 has questioned data).

Column 1 Column 2 Column 3 Column 4 Column 5 Columns 1-4

Number 70 69 69 70 69 278
Chi-Square 8.57 5.93 8.54 9.14 26.22 4.45
D. Freedom 9 9 9 9 9 9
Probability 0.478 0.747 0.481 0.42 4 0.0019 0.880
Table 4. Tests of Uniformity of Digits for the Columns of the Published Table. Chi-Square tests of rightmost digits
for, respectively, Columns 1 to 5 of the Published Table, and for Columns 1-4, together.

notebook data could be found for the questioned

numbers in column 5. Place 4 Place 3 Place 2 Place 1
Of statistical importance is the fact that 1 4 0 4
means and standard deviations in this Table are
7 1 0 7
reported to several places. Thus numbers are
recorded with greater precision than the 9 6 7
repeatability that the biological experiment 1 4 0 7
allows, permitting a digit analysis. 7 1 0 4
The treatment of rightmost digits is the same Table 5. Vertical Pattern of Digits
as that for the previous case. Digits are analyzed
in four places with no leftmost digit included in not done is strengthened.
the analysis. Furthermore, examination of the standard
Only the digits of the questioned Column 5 deviations in the upper half of Column 5 of Table
are significantly different from uniform (p = 3 reveals a remarkable “vertical” pattern. These
.0019). Columns 1 to 4 separately are not numbers should be statistically independent from
different from uniform (the probability ranges row to row. However moving vertically
from .424 to .747). In the aggregate, columns 1 downward at each digit place reveals a
to 4 are again not different from uniform (p = symmetrical recurrence of digits: 1,7, blank, 1, 7;
.88). 4, 1, 9, 4,1; then 0, 0, 6, 0, 0; and finally, 4, 7, 7,
Based on the contrast between the digit 7, 4 (Table 5).
distributions for the questioned Column 5 and the The vertical pattern does not appear
unquestioned columns, the complainant’s consistent with the five presumably statistically
assertion that the experiment for Column 5 was independent experiments depicted by the separate

272
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data

Journal 1 Journal 2 Book determined to occur whenever a

peak on the recording of current
Trauma patients Cancer patients Trauma patients equaled or exceeded 10 picoAmps.
26428 406 6428 406 116428 3406 Since the spontaneous “firings”
7824 376 7824 376 17824 3761 were infrequent, the continuous
record of the electrical signal was
24840 1107 24840 1107 124840 7107
not retained. Instead, an “event
detection” device sent the time and
26660 345 6501 355 116660 34511 the amplitude of the firing to Excel
7791 407 7906 348 17791 407
spreadsheets as a permanent
record of the experiment.
9276 1498 12016 1476 9276 1498 To graph the activity of
Table 6. muscles from different genetic
crosses, the firings of various
rows of Table 3. Such a pattern is consistent with amplitudes were accumulated into bins of 5-
the formation of the numbers after the outline of picoAmp width (10-15, 15-20, 20-25, etc), with
the published Table had been established. accumulation continuing until some bin
Finally, to check for the possible existence of contained 100 firings.6 The resulting frequency
a pattern, three publications by the respondent distribution represented the pattern of firings (for
(two journal articles and a book chapter) were Experiment 1, see Figure 2, below, in which there
examined. Examination of these publications are just over 100 events in the 20-25 bin).
reveals patterns of digits that are inconsistent Prior to publication, the respondent’s
with biological observations. Consider Table 6 coworkers thought that firings should only be
(above), which contains numbers from tables in defined as those peaks 20 picoAmps or greater.
three different publications by the author, all for a Thus they asked the respondent to prepare a new
similar experimental procedure. graph like that of Figure 2, but sampling only
In these three publications, rightmost digits peaks 20 picoAmps or greater (i.e. resampling
that should not be reproducible are the same in the Excel spreadsheet until some bin contained
the first and third rows, and they would be the 100 such firings.)
same in the second row except for the arbitrary The respondent submitted a new frequency
addition of a “1” after the “376” in the last graph that appeared like the first, but truncated at
column. Further, in the fifth row two of the 20 rather than 10. Since one would expect the
standard deviations are “407” while the shape of the new graph (above 20 picoAmps) to
corresponding means are “7791” and “17791.” differ, the coworkers questioned the result.
Note that the standard deviation 7107 occurs in
the book chapter and also in
Column 5 of the published
table already discussed. The
respondent in this case
agreed that the data were
“flawed” and retracted the
relevant articles.

Case 3: Banker’s
rounding and “odd”
terminal digits
For the purposes of a genetic
study, electro-physiological
measurements of
spontaneous “firings” (action
potential spikes) of isolated
muscle fibers were made. Figure 2. Binning of amplitudes into bins of 5-picoAmp width (initial 321
Initially, a firing was records of Experiment 1).
273
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
The respondent asserted that the new graph mid-point of the two is recorded. Thus when
was not simply a truncated version of the first, successive time values are added and divided by
but represented a fresh sampling of peaks greater 2, the resulting terminal digit is 5 and would be
than 20 picoAmps. He asserted that he had rounded to an even digit, for example: (1000 +
properly sampled the peaks in an Excel
1001)/2 = 1000.5 rounds to 1000, and (108.7 +
spreadsheet by counting beyond the initial 321
records on which the first graph (Figure 2) was 108.8)/2 = 108.75 round to 108.8. Therefore if
based. The respondent furnished an Excel numbers ending in 5 are rounded, only even
worksheet, “Experiment 1,” of 551 records in numbers occur. The rounding of terminal 5’s to
support of the new graph. This worksheet the nearest even digit is the ANSI/IEEE standard9
contained the initial 321 records along with 230 for rounding terminal 5’s in computers.
additional records. Examination of the terminal digits of the 1026
In addition to the Excel worksheet for time values of the unquestioned data in
Experiment 1, the respondent also provided a Experiment 2 reveals no times ending in an odd
worksheet of unquestioned data “Experiment 2” digit. (The distribution of the 1026 penultimate
digits of the times for Experiment 2 is not
with 1026 records. For Experiment 1 and the 10
different from uniform (Chi-Square = 14.6, df =
picoAmp peaks, the initial 321 records of 9, p = .10).) In contrast, the questioned
Experiment 1 are largely determined since the Experiment 1 contains time values that end in
initial Figure 2 is known. Thus the last 230 odd digits, reflecting insertions and alterations.
records of Experiment 1 are more questionable. In the initial 321 time points, six terminate in an
Since all 551 records were provided after the odd digit (Figure 3). (The distribution of the 315
allegation, the opportunity existed to falsify or penultimate digits from the potentially unaltered
fabricate time points, but if falsifications occur, even times is not different from uniform (Chi-
most would be expected in the last 230 records. Square = 8.14, df = 9, p = .52).)
Table 7, below, presents the first 12 records of Examination of the graph (Figure 4) of the final
Experiment 1. 230 records of Experiment 1 reveals many more
It is interesting to note that all of the time (58) time values with odd terminal digits10 than
values in Table 7 terminate in an even digit. The Figure 3. (The distribution of the 172 penultimate
occurrence of only even time values can be digits from the even, potentially unaltered, times is
explained by a long-used7 practice sometimes not different from uniform (Chi-Square = 12.3, df =
known as “Banker’s Rounding.8 ” 9, p = .20), whereas the distribution of 58
A simple explanation of the even terminal penultimate digits from falsified times ending in an
digits for time values is that two successive time- odd digit deviates significantly from uniform (Chi-
values are used in determining a peak, and the Square = 33.0, p = .00013).

Figure 3. Experiment 1: first 321 time points; 321 Figure 4. Experiment 1: last 230 time points; 230 terminal
terminal digits from 321 numbers. (Note presence of six digits from 230 numbers. (Note presence of 58 odd digits.)
odd digits.)
274
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data

Experiment 1 - First 12 Records Many more time

Time in Minutes Amplitude in picoAmps Terminal Digit of Time values terminate in odd
0 .0648 16.1 8 digits in the final portion of
0 .4904 22.7 4 Experiment 1, as expected
0 .4952 33.2 2 if falsification occurred.
0 .5398 19.8 8 The occurrence of time
0 .9454 36.1 4 values ending in odd digits,
1 .7182 44.4 2
mostly in the latter part of
2 .6950 20.5 0
Experiment 1 (and the lack
3 .3626 19.3 6
of uniformity of their
3 .7294 17.6 4
penultimate digits) indicates
3 .8586 14.9 6 data falsification. The
4 .3494 12.9 4 timing of the occurrence of
4 .3712 45.4 2 the minutes ending in odd
digits is illustrated in
Table 7. The first 12 records of Experiment 1. Note that amplitudes include values less
than 20, as expected. Also note that the terminal digit of the time is an “even” number Figures 5 and 6.
for all 12 records. From Figure 6 it can
be seen that not only do
most of the odd time
values occur in the last part of
Experiment 1 (after minute
137.3006); it also appears from
the denseness of the plot in the
latter that the values
immediately after this time
point are quite close together.
Further statistical tests of the
intervals between events
confirms the increased density
in the latter part of Experiment
1, indicating the insertion of
fabricated firing events.

Figure 5 (above). Experiment

2, unquestioned, 699 amplitudes
(abs>20). (No amplitude is
associated with an odd
minutes.)

Figure 6 (right). Experiment 1,

questioned; 371 amplitudes with
abs>20, 52 with odd minutes.
(Negative values, even minutes;
positive values, odd minutes.)

275
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Rats
Weights-1 Weights-2 Weights-3 Weights-4 Weights-5 Weights-6
M-1 2.495 3.008 2.7515 4.631 2.250 3.4405
M-2 1.695 2.272 1.9835 3.019 0.702 1.8605
M-3 0.738 1.495 1.1165 1.768 0.843 1.3055
M-4 0.780 0.231 0.5055 0.394 0.085 0.2395
M-5 0.276 0.122 0.199 0.155 0.205 0.180
M-6 4.128 3.413 3.7705 2.261 1.187 1.724
M-7 1.131 1.224 1.1775 2.805 0.726 1.7655

Table 8. Portion of Excel spreadsheet with weights of muscles of rats 1-6. Note that some entries for columns Weights-3
and Weights-6 have four decimal digits and end in 5, whereas other entries have at most three decimal digits.

followed timely procedures.

Case 4: One terminal digit too many
The respondent presented to university
An investigator conducted studies on the effect of
officials blood flow and weight data for six rats
rhythmic contractions of skeletal muscle on
on an Excel spreadsheet as well as purportedly
blood flow using the hind limbs of rats. Blood
original data sheets with handwritten entries for
flow was measured at rest and during nerve
the muscle weights for six rats. Weights of 28
stimulation. In addition to measurements of
muscles and three other body parts for six rats
blood flow, weights of individual skeletal
extracted from the Excel printout are presented in
muscles were recorded on data sheets. The
Table 8 .12 Further weights as found in
experimental results for six rats were presented in
handwritten entries on separate data recording
a laboratory seminar. Sometime later a co-
worker discovered that two of six data sheets sheets for six rats are presented here in Table 9.
were blank, and became suspicious that the In Table 8, columns Weights-1, Weights-2,
measurements (blood flow/weights) had not been Weights-3 and Weights-6 correspond,
made for those rats. Suspicions were confirmed respectively to columns 314-1, 314-2, 315-1 and
when frozen rat carcasses were checked. 316-2 in Table 9. Thus the handwritten
Although four had the hind limb muscles “original” data on the four data recording sheets
dissected, two were still intact and un-dissected. (314-1, 314-2, 315-1 and 316-2) correspond to
When confronted, the investigator (now the columns labeled, respectively, Weights-1,
respondent) admitted to falsifying data for two Weights-2, Weights-3, and Weights-6 on the
experimental animals. However, he subsequently Excel spreadsheet. The columns Weights-4 and
withdrew the admission and denied the charges. Weights-5 do not correspond to two additional
The respondent stated that there was no evidence data recording sheets labeled 315-2 and 316-1.
to support the claims that the research was When values within a spreadsheet are
falsified,11 and that the university had not calculated, rather than transcribed, the numbers
may display more digits of accuracy than the
original numbers that
Rats are the source of the
calculated values.
314-1 314-2 315-1 315-2 316-1 316-2 Therefore, looking for
M-1 2.495 3.008 2 .725 3.859 3.479 3.440 enhanced precision in
M-2 1.695 2.272 1 .984 2.087 1.881 1.861 spreadsheet numbers
M-3 0.738 1.495 1 .117 1.464 1.320 1.306 can indicate that
M-4 0.780 0.231 0 .506 0.269 0.242 0.240 certain numbers have
M-5 0.276 0.122 0 .199 0.202 0.182 0.180 been calculated or
randomly generated by
M-6 4.128 3.413 3 .771 1.933 1.743 1.724
the spreadsheet
M-7 1.131 1.224 1 .178 1.980 1.785 1.766 software.
Table 9. A portion of rat muscles weights from handwritten entries on six data recording Since data are
sheets. Note that all numbers have a precision of three decimal places. presented for six rats,
276
–––––––––––––––––––––––––––––––––––––––––– Mosimann, et al., Terminal Digits and the Examination of Questioned Data

Rat-3 Rat-6
Mean 1,2 Weights-3 315-1 Difference Mean 4,5 Weights-6 316-2 Difference

M-1 2 .7515 2.7515 2.725 0.0265 3.4405 3.4405 3.440 0.0005

M-2 1 .9835 1.9835 1.984 -0.0005 1.8605 1.8605 1.861 -0.0005
M-3 1 .1165 1.1165 1.117 -0.0005 1.3055 1.3055 1.306 -0.0005
M-4 0 .5055 0.5055 0.506 -0.0005 0.2395 0.2395 0.240 -0.0005
M-5 0.199 0.199 0.199 0 0.18 0.18 0.180 0
M-6 3 .7705 3.7705 3.771 -0.0005 1.724 1.724 1.724 0
M-7 1 .1775 1.1775 1.178 -0.0005 1.7655 1.7655 1.766 -0.0005
Table 10. A Portion of the Weights for Rat 3 and Rat 6. The weights for Rat 3 are precisely the means of the respective
weights for Rats 1 and 2. Additionally, the weights for Rat 3 correspond to three decimals to the handwritten weights for
Rat 315-1. (The only exception is the weight for M-1 (shaded) where the rounded 2.752 is transcribed as 2.725.
Correspondingly, the weights for Rat 6 are precisely the means of the respective weights for Rats 4 and 5. Additionally, the
weights for Rat 6 correspond to three decimals to the handwritten weights for Rat 316-2, without exception.
and at most four allegedly were measured, the found on the (presumably) valid sheets labeled
spreadsheet was evaluated for signs that some of 314-1 and 314-2.
the columns contained calculated values, rather Lacking muscle-weight data for two rats, the
than valid data entered from experimental respondent generated weights by twice forming
records. The columns Weights-3 and Weights-6 means of measurements of other rats. The
in the Excel spreadsheet (Table 8) contain a presence of the extra digit in the Excel
number of entries that are recorded to one more spreadsheet provided the needed clue. When the
decimal accuracy than the other columns respondent was shown that the two rats’ weights
(Weights-1, Weights-2, Weights-4, Weights-5). were clearly produced as means, not measures,
Additionally, these same entries for Weights-3 he accepted the finding of scientific misconduct.
and Weights-6 contain one more digit than the
purported original handwritten data as recorded Notes
on the sheets labeled 315-1 and 316-2 (Table 9). 1. 65 Federal Register 76260, December 6, 2000.
This extra precision could not occur from manual 2. A theoretical discussion is found in J. E. Mosimann
entry of the weights from the raw data sheets. and M. V. Ratnaparkhi, “Uniform occurrence of digits
Instead, the presence of an extra digit for folded and mixture distributions on finite
intervals,” Communications in Statistics, 1996, 25(2),
indicates the possibility that these two columns
pp 481-506. Among other issues, this paper discusses
represent calculated data. Further, where the approximations to continuous distributions by
extra digit occurs, it is always a “5.” This histogram-distributions for which the uniformity of
indicates the calculation may have involved terminal digits up to a specified place is known. Such
division by “2,” suggesting that those numbers theoretical issues are important, but our emphasis here
could be the means of two columns. (When the is on direct comparison of questioned data with
sum of the two numbers is even, there is no unquestioned data.
increase of the non-zero digits; however, when 3. On May 1, 1995, the Maryland Lottery initiated a
the sum is odd, division by 2 produces an midday pick-3 drawing for weekdays. This is in
addition to the nightly drawing. Thus there are more
additional “5” digit.)
than 3,650 winning pick-3 numbers over the ten-year
In fact, the column Weights-3 is precisely the period. Maryland Lottery results may be found at the
mean of columns Weights-1 and Weights-2 (see official website, http://www.mdlottery.com.
Table 10, below). Correspondingly, the column 4. In all there are 474 counts: 252 admittedly falsified
Weights-6 is the mean of columns Weights-4 and (notebook pages 141-152) and 222 unquestioned
Weights-5 (Table 10). counts that are supported by counter printouts
Since these two columns are calculated on (notebook pages 104-106, 130-131, 134-135). Each
the spreadsheet, the “original” data on the count, falsified or unquestioned, contains from three to
six digits. Digits were tested in four places, but no
recording sheets 315-1 and 316-2 are copied,
digit that was itself the leftmost digit was included in
respectively, from the spreadsheet-calculated the analysis. Total analyses included 939 digits from
columns Weights-3 and Weights-6. The only 252 falsified numbers and 857 digits from 222
modification is that the “original” copied data are unquestioned numbers.
only transcribed to three-decimal accuracy as
277
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
5. See “Data Fabrication: Can people generate Random
Digits?” J. E. Mosimann, C. V. Wiseman and R. E.
Edelman, Accountability in Research, 4, 31-55, 1995.
This study shows that many people have difficulty
fabricating random digits, even when trying to do so.
6. “Inverse” sampling until a certain number of a
particular event occurs has a long history, particularly
where rare events are to be studied. (For example, see
J. E. Mosimann, “On the compound negative
multinomial distribution and correlations among
inversely sampled pollen counts,” 1963, Biometrika,
50, 47-54).
7. “It is conventional to round off to the nearest even
digit when the number to be rounded is exactly half
way between two successive digits.” pp. 13-14, Paul
S. Dwyer, Linear Computations, 1951, John Wiley &
Sons Inc., i-xi , 1 – 344. (See also the next two
footnotes.)
8. “PowerBASIC always rounds towards the closest even
number. For example, both 1.5 and 2.5 would be
rounded to 2. This is called banker’s rounding. …” p.
169, User’s Guide, 1997, PowerBASIC, Inc. 316 Mid
Valley Center, Carmel, California, i-vi, 1-318.
9. ANSI/IEEE Std 854-1987, October 5, 1987, “ANSI”
denotes the American National Standards Institute and
“IEEE” denotes the Institute of Electrical and
Electronic Engineers, Inc. “4.1 Round to Nearest. …if
the two nearest representable values are equally near,
the one with its least significant digit even shall be
delivered.” “5.4 Round Floating Point Number to
Integral Value. …when rounding to nearest, if the
difference between the unrounded operand and the
rounded result is exactly one half, the rounded result is
even.”
10. 46 of these 58 time values that terminate in odd digits
occur with amplitudes greater than 20 picoAmps. In
the initial 321 records of Experiment 1, 6 of 6 odd
time values occur with amplitudes greater than 20
picoAmps.
11. It is only after the respondent denied the charges and
findings of the institution that the ORI demonstrated
which two rats on the spreadsheet represented falsified
data, and the manner of falsification.
12. The spreadsheet also contains columns of numbers
representing blood pressure measurements and
radioactive counts, some of which the university
committee regarded as falsified. These are not
presented here.

278
9. Publication Practices
Guidelines on Plagiarism and Paraphrasing in Writing Manuals Across
Various Disciplines
Miguel Roig, Department of Psychology, St. John’s University, NY, USA

Jaclyn de Jacquant, Department of Psychology, St. John’s University, NY, USA

Keywords: Paraphrasing, Plagiarism, Writing guides

Discussions of plagiarism in conventional writing manuals typically focus on acknowledging the

source of borrowed ideas and text. Such coverage often includes guidelines for proper attribution and
citation practices. A number of manuals also provide specific guidelines for correct paraphrasing. By
correct paraphrasing, we mean the extent to which text from an original source should be modified in
order for it not to be considered a potential case of plagiarism. Those manuals that cover proper
paraphrasing practices (1-3), generally suggest that, in addition to providing a citation, authors should
always paraphrase others’ work using their own words and expressions and avoid the use of the
original author’s language. For example, in a widely used guide, the authors state “When you
paraphrase or summarize, you should use your own words and sentence structure (4). Imitating
syntax, rearranging words and phrases, and borrowing phrases even as brief as two or three words do
not change the original sufficiently to avoid plagiarism” (pg. 66).
Aside from the above guideline on paraphrasing, we are not aware of any other major writing
manual that provides as close an operational definition for correct paraphrasing as the above example
illustrates. However, the examples of proper paraphrasing provided by conventional manuals that
offer such coverage suggest that a correct paraphrase must represent a very substantial modification
of the original text, otherwise the paraphrase may constitute plagiarism. Moreover, some manuals
such as the one quoted above, even suggest that, to avoid plagiarism when paraphrasing, not only
should the original words be changed, but also the sentence structure of the newly paraphrased text
must be different from that of the original (4-7).
As the reader might suspect, the criteria for correct paraphrasing appear to differ from writer to
writer, particularly for inexperienced writers. For example, recent studies by one of the present
authors have reported wide differences in plagiarism/paraphrasing criteria among college students (8,
9). Furthermore, similar differences also appear to exist among professionals, including physicians,
English professors, and journal editors, and between college professors from a variety of disciplines
(10-11). Some authors have even begun to express concern about the writing practices of those who
engage in ‘light’ paraphrasing of others’ works and terms, such as ‘patchwriting’ and
‘paraphragiarism’, have been offered to describe some of these inappropriate paraphrasing practices
(12-14).
Depending on a number of factors, federal agencies, such as the National Science Foundation and
the Office of Research Integrity do not classify inappropriate paraphrasing as instances of research
Corresponding author: Miguel Roig, Department of Psychology, Notre Dame Division of St. John’s College, St. John’s
University, 300 Howard Avenue, Staten Island, NY 10301, 718-390-4513 (voice), 718-390-4347 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
misconduct (15). However, based on definitions 1). First, we proceeded to determine each
provided by conventional writing manuals and, manual’s extent of coverage of plagiarism by
depending on the context, others may still judge reviewing its index and table of contents for
such inappropriate writing practices as potential entries for ‘plagiarism’ and for ‘paraphrasing’. If
instances of plagiarism. Thus, the ‘light’ no entries were found for those terms we
paraphrasing of others’ text, an innocuous writing proceeded to examine sections on citation and
practice to some, can have serious consequences documentation procedures.
and possibly result in disciplinary actions by the
individual institutions and/or the academic Results
disciplines involved. Most of the manuals were found to provide some
A matter that we believe to be of major discussion of citation and quotation procedures.
concern is evidence that suggests that Indeed, these sections are designed primarily for
inappropriate paraphrasing practices on the part the purpose of identifying the source of ideas and
of academics may be much more common than thus, prevent an interpretation of plagiarism.
most people assume. For example, in a recent Surprisingly, only 3 of the writing manuals
series of studies (11), one of the present authors examined (1, 17-18) listed entries for plagiarism
found substantial differences in paraphrasing in their index. The extent to which plagiarism
criteria among college professors from a variety was covered in these three sources varied
of disciplines, including professors in the somewhat. All three manuals provided some
sciences. In one of the studies, differences in discussion of plagiarism. But, only two, the
paraphrasing criteria arose even among members Modern Language Association (MLA) manual
of a single academic discipline: Psychology. (1) and the American Medical Association
These findings led the author to review the (AMA) manual (18) defined this type of
official guidelines for avoiding plagiarism transgression and provided specific examples of
published by the American Psychological instances of plagiarism (e.g., word for word
Association (APA), the parent association of lifting of a passage without attribution;
psychologists (16, 17). A close examination of presenting others’ ideas without attribution).
these guidelines revealed a certain degree of The two writing guides that included
ambiguity in how correct paraphrasing and coverage of paraphrasing (17, 18) defined it as
plagiarism are defined in that discipline. That restating text in the author’s original words, but
particular finding is noteworthy because one of only the APA manual (17) provided an example
the sources reviewed is not only used by of proper paraphrasing. However, as one of the
psychologists, but also by members of other present authors has pointed out, the definition for
disciplines (e.g., sociology and education) (17). paraphrasing provided by the APA (i.e.,
Given the importance of avoiding plagiarism “Summarizing a passage or rearranging the order
in scholarly and scientific writing, the above of a sentence and changing some of the words is
findings raise a number of important questions: paraphrasing.”) appears to be somewhat at odds
How do other disciplines in the sciences and the with the actual example offered (11). That
humanities define plagiarism? What are their example, which shows the original text to have
guidelines regarding correct paraphrasing? How been substantially modified, is consistent with
similar are these definitions across disciplines? other conventional manuals’ examples
In an attempt to address these questions, we paraphrasing.
surveyed the writing manuals of various
disciplines within the sciences and humanities for Discussion
their coverage of plagiarism. We were interested Given the importance of avoiding plagiarism, we
in the extent to which definitions of plagiarism, are somewhat concerned with the fact that the
specifically guidelines for correct paraphrasing, writing manuals of several academic disciplines,
are covered in these manuals and the degree to particularly many disciplines in the sciences, do
which such definitions are consistent across not appear to have explicit sections on these
disciplines. matters. We note that other important resources
on writing in the humanities and in the
Method biomedical sciences also appear to lack entries on
We located the latest edition available to us of plagiarism (19-21).
writing manuals of various disciplines (Appendix It is possible that at least some of these
282
–––––––––––––––––––––––––––––––––––––––––––––––––– Roig & de Jacquant, Guidelines on Plagiarism and Paraphrasing
manuals do provide some coverage of plagiarism. Bibliography
But, in addition to not listing the term 1. Gibaldi J. MLA Style Manual and Guide to Scholarly
‘plagiarism’ in the manuals’ index or table of Publishing. 2nd ed. Modern Language Association of
contents, any coverage, if it occurs, is probably America; 1998.
very minor at best and takes place in sections 2. Kirszner LG, Mandell SR. The Holt Handbook. 5th ed.
Fort Worth: Hartcourt Brace; 1999.
other than those we reviewed.
3. Nadell J, McMeniman L, Langan J. The MacMillan
If most of these manuals do not provide Writer: Rhetoric and Reader. 2nd ed. New York:
coverage of plagiarism the reason may be an MacMillan; 1994.
assumption on the part of authors and editors of 4. Campbell WG, Ballou SV, Slade C. Form and style:
these reference materials that contributors to the Theses, reports, term papers. 8th ed. Boston: Houghton
professional literature are already knowledgeable Mifflin; 1990.
about such fundamental matters of scholarship. 5. Aaron JE. The Little, Brown Compact Handbook. 3rd
Indeed, some manuals written for students in ed. New York: Longman; 1998.
disciplines, such as biology, psychology, and 6. Hacker D. Writer’s Reference. 4th ed. Boston, Bedford;
1999.
sociology provide definitions of plagiarism and
7. Troyka LO. Simon & Schuster Handbook for writers.
paraphrasing that are consistent with those of 5th ed. Upper Saddle River, NJ: Prentice Hall; 1999.
conventional writing manuals that provide 8. Roig M. Can undergraduate students determine whether
coverage of these issues (22-25). Such detailed text has been plagiarized? The Psychol Record
coverage at the undergraduate level supports the 1997;47:113-122.
assumption that, at the professional level, authors 9. Roig M. When college students’ attempts at
already know the rules. Another reason for not paraphrasing become instances of potential plagiarism.
including coverage may be that, as an ethical Psychol Rep 1999;84:973-982.
issue, plagiarism is likely to be addressed in 10. Julliard K. Perceptions of plagiarism in the use of other
author’s language. Fam Med 1994;26:356-360.
other sources of information, such as a
11. Roig M. Plagiarism and paraphrasing criteria of college
discipline’s code of ethics. Finally, sections on and university professors. Ethics and Behav 2001;
citation procedures represent, to a great extent, a 11:307-323.
discipline’s way of insuring that authors of 12. Howard RM. Plagiarisms, authorships, and the
original works are properly credited. Therefore, academic death penalty. College English 1995;57:788-
although explicit sections on plagiarism might 806.
not be provided in many of the writing guides 13. Howard RM. The new abolitionism comes to
reviewed, there is an implicit message in these plagiarism. In: Buranen, L, & Roy LM editors.
guides that authors must duly credit others whose Perspectives on plagiarism and intellectual property in a
postmodern world. NY: State University of New York;
ideas, text, or data are being borrowed.
1999. p. 87-95.
In spite of the above considerations, and in 14. Levin JR, Marhsall H. Publishing in the Journal of
view of the fact that plagiarism continues to Educational Psychology: Reflections at Midstream
flourish, we believe that writing manuals across (Editorial). J Educational Psychol 1993;85:3-6.
all disciplines should provide explicit sections on 15. Parish D. Scientific misconduct and the plagiarism
plagiarism that include clear definitions and cases. J College and University Law 1995;21:517-554.
examples of the various forms that plagiarism 16. American Psychological Association. Ethical principles
can take. In addition, given that a significant of psychologists and code of conduct. Am Psychol
portion of scholarly writing involves 1992;47:1597-1611.
17. Publication Manual of the American Psychological
summarizing and paraphrasing others’ ideas and
Association. 4th ed. Washington, D.C. American
text, writing manuals should pay particular Psychological Association; 1994.
attention to this area and offer clear guidelines as 18. Iverson C. et al. American Medical Association Manual
to what forms of writing constitute proper of Style. A Guide for Authors and Editors. 8th ed.
summarizing and paraphrasing techniques. Baltimore: Williams and Wilkins; 1998.
Finally, and perhaps most difficult of all, 19. The Chicago Manual of Style. 14th ed. Chicago:
definitions of plagiarism and guidelines for University of Chicago Press; 1993.
summarizing and paraphrasing text should be 20. Reif-Lehrer L. Grant application writer’s handbook.
standardized across disciplines. We believe that, Jones & Bartlet; 1995.
21. Uniform requirements for manuscripts submitted to
in the absence of such standardization and given
biomedical Journals. Ann Intern Med 1997;126:36-47.
the increasing nature of cross-disciplinary 22. McMillan VE. Writing papers in the biological sciences.
collaborations, there is the potential for an even 2nd ed. Boston: Bedford/St. Martin’s; 1997.
greater number of plagiarism cases in the future.
283
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
23. Pechnik JA. A short guide to writing about biology.
Boston: Little Brown & Company; 2000.
24. Szuchman LT. Writing with style: APA style made easy.
Pacific Grove, CA: Brooks/Cole; 1999.
25. Richlin-Klonsky JR, Stresnki E. A guide to writing
sociology papers. 4th ed. New York: St. Martin’s Press;
1998.

Appendix 1: Writing manuals reviewed

for their coverage of plagiarism and
paraphrasing.
American Institute of Physics. AIP Style Manual. 4th ed.
New York: AIP; 1990.
American Mathematical Society. A Manual for Authors of
Mathematical Papers. Revised ed. Providence: AMS;
1990.
Associated Press Staff. Associated Press Stylebook and
Libel Manual. Reading, MA: Addison; 1998.
Council of Biology Editors. CBE Style Manual: A Guide
for Authors, Editors, and Publishers in the Biological
Sciences. 5th ed. Bethesda: CBE; 1994.
Dodd JA. Editor. The ACS Style Guide: A Manual for
Authors and Editors. Washington: American Chemical
Society; 1998.
Gibaldi J. MLA Style Manual and Guide to scholarly
writing. 2nd ed. New York: Modern Language
Association of America; 1998.
Holoman DK. Writing about Music: A Style Sheet from the
Editors of 19th-Century Music. Bekeley: U of
California; 1988.
Iverson C. et al. American Medical Association Manual of
Style. A Guide for Authors and Editors. 8th ed.
Baltimore: Williams and Wilkins; 1998.
Linguistic Society of America. “LSA Style Sheet”. LSA
Bulletin 1999 Dec; 166: 64.
Publication Manual of the American Psychological
Association. 4th ed. Washington, D.C. American
Psychological Association; 1994.
The Bluebook: A Uniform System of Citation. 16th ed.
Comp. Editors of Columbia Law Review et al.
Cambridge: Harvard Law Review; 1996.
The Chicago Manual of Style. 14th ed. Chicago: University
of Chicago Press; 1993.

284
Instructions to the Author: An Integrity Issue
Mary D. Scheetz, Office of Research Integrity, DHHS, USA

Keywords

Corresponding author:
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

286
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Scheetz, Instructions to the Author

287
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

288
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Scheetz, Instructions to the Author

289
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

290
Erratum Citation and Accuracy in the Publication Record
Marshall Thomsen, Department of Physics and Astronomy. Eastern Michigan University,
USA
Christopher Aubin, Department of Physics and Astronomy. Eastern Michigan University,
USA
Barbara Hall, Department of Physics and Astronomy. Eastern Michigan University, USA
Matthew Knee, Department of Physics and Astronomy. Eastern Michigan University, USA

Keywords: Erratum, Citation, Publication

Technological advances have greatly influenced the ways in which science is communicated.
However, the refereed journal remains an important element of the system, providing a permanent
record of information with some quality control over the scientific content. In trying to keep abreast
of recent developments in a field or when entering a field of study for the first time, scientists often
rely on the refereed journal as their primary information source. Thus accuracy of the written record
becomes a significant issue.
While much has been written about the publication process in general, (1) we will focus on a
small piece of the process that lends itself to accumulation of basic statistical information and, we
hope, provides some insight into other broader aspects of publication. In particular we will look at
physics papers that have an erratum associated with them and study how these papers are cited in
subsequent literature. There are several issues we will examine. If an erratum is written, how likely
is it that those who have read the original paper also will have read the erratum? If a corrected paper
is cited, how likely is it that the authors who cited the paper also cited the erratum? Is it misleading to
cite the original paper but not the erratum? Do authors typically cite their own errata?
Some of these questions have been addressed before. For instance a 1990 study of retracted
medical papers showed that retractions tended to reduce, but not eliminate, citation rates. (2) A 1995
study of errata in physics journals showed that when corrected papers are cited, most often the
corresponding erratum is not cited. (3) The authors of the study commented at the time that part of
this citation problem was associated with the logistical issue of locating an erratum. It is much easier
to search the publication record backward in time by studying citations. Moving forward in time to
locate errata requires scanning journal contents or using an index (such as the Science Citation Index).
The authors speculated that as more journals were provided in an electronic format, locating errata
would be easier since the original paper presumably would be linked electronically to the erratum.
The American Physical Society now has a large collection of its journals available online via
subscription. All of their recent online papers that have an associated erratum have a link to that
erratum. We thus undertook a new study to determine if this electronic linking has improved the
Corresponding author: Marshall Thomsen, Department of Physics and Astronomy, Eastern Michigan University, Ypsilanti,
MI, 48197, 734-487-8794 (voice), 734-487-0989 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
citation rate of errata. Curiously, we find that, if primary papers and/or their errata. We refer to
anything, the citation rate for errata seems to this collection of papers as secondary papers. It
have decreased since the introduction of the is interesting to note that a small portion of these
electronic format. secondary papers cited only the erratum and not
the corresponding primary paper. As a spot
Study Design check on the accuracy of Science Citation Index,
Our study consisted of an examination of the we used the citation list provided by the online
citations of 14 papers from Physical Review version of Physical Review Letters. It should be
Letters in 1995 and 1996 that had errata making noted that the journals used in this citation index
nontrivial corrections. We included papers with are much more limited in scope than those used
calculational errors requiring replotting graphs or to assemble the Science Citation Index, listing
reproducing tables, papers in which derivations citations by only American Physical Society
needed modifications, papers in which data journals. We selected three primary papers
needed to be reanalyzed due to misinterpretation, from our list that, according to the Science
etc. We excluded papers in which simple Citation Index, had no citations to their erratum.
typographical errors were corrected or We verified this finding with all available listings
acknowledgments of related work were added. on the more limited Physical Review Letters
The goal was to focus on papers in which there citation data base and also confirmed that all 21
was a scientific error of substance being secondary papers appearing on this database also
corrected by the erratum. At the same time, none appeared on the Science Citation Index data base.
of the errata reported on here amount to a That is, we discovered no evidence that Science
complete retraction of a paper. For clarity in Citation Index was omitting papers appropriate
discussions below, we refer to these 14 papers as for our secondary category.
the primary papers.
We selected our primary papers from Results and Discussion
amongst the first papers to become available in The collection of secondary papers was divided
the Physical Review Letters online collection. into two categories. The first category contained
Hence the primary papers all have electronic those papers in which there was an overlap
links to their errata, and they have all been between the authors of the secondary paper and
available in the literature for several years (thus those of the cited primary paper. The second
increasing their chances of being cited). category consisted of those secondary papers in
Physical Review Letters is one of the most which there was not any overlap of authorship
selective physics journals, containing papers with the cited primary paper. The purpose of this
describing some of the most recent and division was to address separately the questions
significant advances across all physics of how often authors cite their own errata and
disciplines. We focussed on these papers since how often independent authors cite errata. The
they are higher profile and hence likely to cases with overlapping authors will be
produce a greater set of citation data. In contrast, considered first.
the 1995 study of errata in physics journals Table 1 shows data for authors citing their
included papers from both Physical Review own errata. We exclude from consideration in
Letters and Physical Review B, the latter being a the secondary paper data set those papers
more specialized journal. That study showed that published prior to the appearance in print of the
papers in Physical Review Letters tend to be cited erratum. We are left with 59 secondary papers
two to three times as often as papers in Physical that could have cited an erratum. Of these, 25
Review B. The 1995 study also showed the (42%) actually did cite the erratum. The reason
citation rate for errata in Physical Review Letters for the remaining 58% of the secondary papers
was substantially higher than that for Physical not including the erratum citation is not clear.
Review B. Thus our present study focuses on a One possibility is that the author of the primary
journal with a relatively high erratum citation paper and erratum chose not to cite the erratum.
rate. Another possibility is that the person or persons
We attempted to identify all papers that had of the secondary paper who took the most
cited the primary papers and/or their associated responsibility for writing that paper were not
erratum, using the Science Citation Index as our among the authors of the primary paper. In this
main tool. We located 507 papers citing the case, it would be possible for the writer of the
292
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Thomsen et al., Erratum Citation and Accuracy

Paper Potential Actual average from the previous study. It is thus safe to
Identification Erratum Erratum conclude that the advent of electronic journals
has not had the desired impact on erratum
Number Citations Citations citation.
1 1 1 We now return to the issue of the extent to
2 5 0 which it is a problem that errata are not generally
3 2 0 being cited. There are three fundamental
4 4 2 questions. First, does the reader of the secondary
5 4 3 paper need to be aware of the erratum? Second,
will a reader discover an erratum based on
6 3 2 information provided by the authors of a
7 0 0 secondary paper? Third, whose responsibility is
8 5 2 it to locate the erratum?
9 3 1 We will examine the first question in the
10 1 0 context of the errata discussed here: those
11 5 0 providing substantive corrections. The 1995
study of erratum citations showed that a little
12 3 0 more than half of the primary papers examined
13 23 14 were cited “in passing” in the secondary
14 0 0 reference. In these cases, the secondary authors
Total 59 25 were primarily acknowledging the work of others
in the field rather than laying down specific
Table I: Analysis of citations by one or more authors of the ground work for their own paper. These citations
original (corrected) paper. Potential erratum citations typically occur in the introductory section. The
represent the total number of papers citing the original remaining citations to the primary papers
paper, its erratum, or both. Actual citations represent the
number of times the erratum was cited. Only papers
appearing after the publication date of the erratum were Paper Potential Actual
considered in columns 2 and 3. Identification Erratum Erratum
secondary paper to be unaware of the existence Number Citations Citations
of the erratum. However, assuming the erratum 1 6 0
author read through the secondary paper prior to 2 4 0
publication, then either that author chose not to 3 22 0
add the erratum citation to the list or overlooked 4 13 8
the absence of the erratum in the references. We
will return to this issue later. 5 8 2
Table 2 shows data for secondary papers 6 15 7
sharing no authors in common with the cited 7 3 1
primary paper. We exclude from the secondary 8 6 6
paper data set those papers that did not appear in 9 2 0
print at least one year after the publication date of
the erratum. This is to ensure that the authors of
10 2 2
the secondary paper had the chance to see the 11 8 0
erratum at the time they were writing their own 12 17 1
paper. After reducing the data set as described, 13 248 32
355 secondary papers remained. Of these, just 14 1 0
59 (17%) cited the erratum. The 1995 study of 9 Total 355 59
primary papers in Physical Review Letters had a
citation rate of 39% (51 of 131) when a similar Table II: Analysis of citations not involving authors of the
original (corrected) paper. Potential erratum citations
approach to data analysis was used. While there
represent the total number of papers citing the original
are obviously statistical fluctuations associated paper, its erratum, or both. Actual citations represent the
with this sampling, it is worth noting that only 4 number of times the erratum was cited. Only papers
of the 14 primary papers in the present study had appearing one year or more after the publication date of the
an erratum citation rate exceeding the 39% erratum were considered for columns 2 and 3.

293
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
indicated that the authors of the secondary paper This brings us to our final question: Who is
were using one or more results or ideas from the responsible for locating the erratum? It is
primary paper to support their own work. This reasonable to view a reference to a paper as a
latter group of citations raises the erratum recommendation of a source to consult for further
citation question in a direct way. Even if the information. In making that recommendation, an
erratum did not have any direct bearing on the author thus has some responsibility to ensure that
portion of the primary paper that was drawn it is a sound recommendation. However, a reader
upon, citing the erratum is still significant in that of a secondary source who is making an in depth
it indicates that the secondary authors are aware study that requires consulting cited references
of its existence and took it into account (if also bears some responsibility for seeking out
necessary) in preparing their paper. Furthermore, relevant errata. While it is difficult to say who
a reader of the secondary paper who is inclined to has the greater responsibility, neither side can be
investigate the topic more thoroughly can be removed from the equation.
misled if unaware of the existence of the erratum. It is worth noting that the secondary author is
Returning to the citations “in passing,” there somewhat more likely to be aware of the erratum
are typically two motivations for providing such than the reader of the secondary paper, because
a citation. First, one may wish to pay tribute to often one cites papers written by people with
predecessors in a particular field. Second, one whom one has had some direct or indirect
may wish to direct the reader to papers with association or by people whose work one has
relevant background information. Papers cited followed closely. This correlation of course is
for the second reason also should have their particularly true in the case of a secondary author
corresponding errata cited as a service to the also being a primary author. This observation
reader. coupled with the fact that erratum citation is not
We now consider the second question: Will a routine even when there is an overlap between
reader discover an erratum based on information primary and secondary authors leads us to
provided by the authors of a secondary paper? speculate that secondary authors are not always
Obviously, if the authors have cited the erratum, citing errata even when they are aware of their
the answer is yes. If the authors have not cited existence. Why is this the case? One possible
the erratum, then there are a number of ways in argument is that some perceive there is a stigma
which the reader may discover the erratum. For associated with publishing an erratum and hence
instance, the reader may look up the primary they prefer not to call attention to it. Arguably,
paper electronically and discover a link to the however, publishing an erratum is a sign of both
erratum. This is constrained by the fact that not integrity and attention to detail. It is likely most
everyone has access to journals in electronic physicists who have done any significant amount
form and not all journals are available in this of research have encountered papers that should
format. When using journals in printed format, have had errata but the authors chose not to write
the reader must rely on techniques such as one. Clearly there is more damage to the field by
searching the journal index for errata or using a uncorrected papers than by those properly
more comprehensive index such as the Science corrected. The irony is that if one takes the time
Citation Index. Otherwise, the erratum might to do the right thing—to write the erratum—it is
only be discovered by chance while browsing not clear how many people are going to read it.
through an issue.
Perhaps authors of the secondary papers Conclusions
assume that the interested reader will be able We conclude as the previous study did with the
locate errata on their own. Short of taking a hope that eventually the conversion of printed
survey, we can only speculate as to whether this journals into electronic databases will resolve the
is the rationale for authors not citing errata. erratum citation problem. In particular, if we
However, given the fact that this citation problem reach a point where all journals are in a dynamic
predates the electronic journal format, it is electronic database that is updated with
unlikely that most authors are consciously appropriate links as errata are written and
electing not to cite an erratum on these grounds. electronic access is as pervasive as printed
It is possible, however, that this rationale may access, then it becomes unnecessary to cite
explain the drop in the erratum citation rate errata. While many physics journals are headed
between the 1995 study and the present study. in this direction, it is not clear if and when all
294
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Thomsen et al., Erratum Citation and Accuracy
will get there. Particularly problematic is the Paper Paper Erratum
task of going through older journals and Number Reference Reference
converting them to electronic format. In the
1 74:694 75:355
meantime, citing errata will continue to be an
important part of the service provided by authors 2 74:1839 76:4097
in their reference sections. 3 74:4101 76:4293
Even if the erratum citation problem is
resolved, the fact that it has existed raises more 4 75:1447 75:3781
general questions concerning the integrity of the 5 75:394 75:1874
publication record. Specifically, is the accepted
6 75:3549 77:2345
norm that authors do have a responsibility to cite
errata or is the expectation that the reader is 7 75:4413 76:3242
responsible for locating them? More generally, is 8 76:014 76:2826
this problem a sign of pervasive sloppy practices
in publication or is it merely a situation of ill- 9 76:1031 77:4278
defined responsibility? The answers to these 10 76:2848 77:5148
questions will become clearer only after more
11 76:3955 78:3227
discussion within the scientific community.
12 77:127 78:3587
Acknowledgments 13 77:3865 78:1396
It is a pleasure to acknowledge support from the
Graduate School, the College of Arts and 14 77:4066 78:162
Sciences, and the Department of Physics and
The table above provides references to the papers used in
Astronomy, all at Eastern Michigan University. this study. All are from Physical Review Letters, published
by the American Physical Society. The format is
Bibliography volume:beginning page.
1. See, for instance, LaFollette MC. Stealing into print:
fraud, plagiarism, and misconduct in scientific
publishing. Berkeley (CA): University of California
Press; 1992.
2. Pfeifer MP, Snodgrass GL. The continued use of
retracted, invalid scientific literature. Journal of the
American Medical Association 1990 March 9; 263(10):
1420-1423.
3. Thomsen M, Resnik D. The effectiveness of the erratum
in avoiding error propagation in physics. Science and
Engineering Ethics 1995; 1(3):231-240.

295
10. Theory and Models from other Disciplines
An Epistemic Model for Moral Hazards in Scientific Enterprises
Jean Maria Arrigo, Irvine, CA, USA
Maj-Britt Juhl Poulsen, Department of Medical Philosophy and Clinical Theory, University
of Copenhagen, Denmark

KEY WORDS Adversarial epistemology, Biomedical research, Cooperative epistemology, Dynamics of

competition and cooperation, Epistemic model, Ethics of science, Moral hazards in scientific enterprises,
Political and military intelligence

The immediate connection between successful science and ethical science is weak, so any urgency for
successes may invite ethical lapses. We present a model of the dynamics between methods and
morals in scientific enterprises. The developmental course of scientific enterprises generates
characteristic moral hazards and opportunities, as we exhibit in our case study of a collaboration
between two biomedical research teams. Lastly, we argue that our model offers conceptual gains in
unifying “ethics of research” and “ethics of application” (1, p. 503) and offers practical gains in
guiding codes of science ethics.
Interviews with biomedical researchers (2) and with military intelligence professionals, together
with archived oral histories of weapons researchers, underlie our model (3). Reviews by military
intelligence interviewees improved it iteratively.

A Model of Methods and Moral Hazards in Scientific Enterprises

In our model, the 17th Century Enlightenment vision of science constitutes the prototype of a
cooperative epistemology (theory of knowledge). Epistemic partners freely share targets of
inquiry, observations, and analyses. Empirical inquiry generates layers of knowledge through:
(a) observation of a phenomenon, (b) analysis of observations, (c) meta-analysis of analyses, and so
on.
Political and military intelligence, in contrast, constitutes the prototype of an adversarial
epistemology. Epistemic adversaries may conceal targets of inquiry, observations, and analyses, and
may spy on or sabotage the Adversary’s inquiries. Empirical inquiry by agent and adversary generate
interleaved layers of knowledge through: (a1) Agent’s investigation of a phenomenon (e. g.,
Manhattan Project study of nuclear fission), (a2) Adversary’s possible investigation of the
phenomenon (e. g., possible Japanese study of nuclear fission in World War II), (b1) Agent’s
investigation of Adversary’s possible investigation through espionage, (b2) Adversary’s investigation
of Agent’s possible espionage, and so on. Each investigation by Agent or Adversary includes all the
processes of observation and analysis in cooperative investigation above—and is often performed by
an epistemic subcontractor, such as a scientist or historian, with cooperative methods. The
Corresponding author: Maj-Britt Juhl Poulsen, Department of Medical Philosophy and Clinical Theory, Build. 22.3,
University of Copenhagen, Panum Institute, Blegdamsvej 3, DK-2200 Copenhagen N, Denmark, +45 3532 7596 (voice),
+45 3532 7938 (fax), [email protected]; for reprints: Jean Maria Arrigo,110 Oxford St., Irvine, CA 92612,
[email protected] .
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Adversarial Epistemology Cooperative Epistemology

I. Partisanship: the goal of inquiry is conscious, I'. Impartiality: the goal of inquiry is
strategic advantage over an Adversary. knowledge per se or its nonpartisan utility.
II. Deceptiveness of phenomena: all observations II'. Accessibility of phenomena: the natural
are vulnerable to deliberate deception by the world is not inherently deceptive (René
Adversary, whether by omission or commission. Descartes’ premise).
III. Urgency: the Adversary is dangerous and III'. Deliberation: method leads to superior
implacable so decision is urgent in the short run. results in the long run (Charles Peirce’s
“self-corrective” hypothesis).
IV. Subordination: researchers’ clients govern the IV'. Independence: researchers themselves
broad topics, opportunities, and constraints of govern the topics and methods of inquiry.
inquiry.
Table 1. Poles of the epistemic continuum
adversarial epistemology is thus far more social constructionist paradigm occupy
complex than the cooperative epistemology. The intermediate positions.
complexity encourages intuitive or artful Four principles separate the most extreme
approaches in attributing intentions and positions of the adversarial epistemology from
meanings to the adversary’s behavior. Our the corresponding principles of the most extreme
formulation contrasts the rational basis of positions of the cooperative epistemology, as
military and political intelligence with the stated in Table 1 (5).
rational basis of science. In the adversarial epistemology, deception by
A recent headline story of scientific the adversary leads to secrecy,
misconduct illustrates the interleaving layers of compartmentalization of knowledge, reward of
adversarial investigation. In 1981 the researchers on the basis of loyalty as well as
archeologist Shinichi Fujimura (Agent) ability, and organizational structures that limit the
unearthed the oldest artifacts in Japan, 40,000 scope of inquiry. Repeated use of any research
years old. Some critics (Adversary) became technique or conceptual schema offers the
skeptical of his celebrated discoveries, which by adversary an opportunity for sabotage, which
1993 had pushed the origins of Japanese raises the value of innovation over perfection.
civilization back 700,000 years. Mindful of their Urgency creates trade-offs between accuracy and
suspicions, Fujimura surreptitiously buried utility. Fear of surprises from the adversary
artifacts that he later “discovered” in the presence promotes crude study of broad fields in
of witnesses. Journalists documented with preference to fine study of narrow fields.
hidden video cameras his burials of artifacts. Researchers’ subordination to decision makers
Anticipating Fujimura’s defenders, the journalists creates a distinction between the complex pattern
filmed Fujimura’s fraud at a second site before “knowledge” held by researchers and the
exposing him. Aware of the limitations of the simplistic linear “information” provided to
journalists’ investigations, Fujimura denied clients for decision making.
planting artifacts at sites he had excavated Consideration of typical epistemic
previously. Japan’s Cultural Affairs Agency, now adversaries in science-related enterprises
doubting Fujimura, plans reviews of his earlier suggests the pervasiveness of adversarial
excavations. Critics speculate that Fujimura’s epistemic methods, as indicated in Table 2.
subterfuge may have set back Japanese
archeology a decade (4). A Case Study of Competition and
Cooperation among Biomedical Teams
The Epistemic Continuum Biomedical research can be described as
The adversarial and cooperative prototypes stand collective research, for cooperation among
as opposite poles on a continuum of epistemic individuals is necessary to reach the goals of
commitments. Cosmology and plant taxonomy research. At the same time, biomedical
lie towards the cooperative pole. Biological researchers need credit for their work and
warfare research and forensic psychiatry lie discoveries to make a career and to bring their
towards the adversarial pole. Biometrics, clinical own research programs to fruition. In this way
trials, educational testing, and research in the the interplay of cooperation and competition is an
300
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Arrigo & Poulsen, Epistemic Model for Moral Hazards
Domains of Common Epistemic Adversaries Historical Prototypes
Inquiry of Researchers
Basic sciences Colleagues, rivals, proponents of conflicting paradigms, ethics Watson & Crick,
committees, institutional authorities, peer reviewers, funding The Double Helix
agencies
Medical Institutional Review Boards, regulatory agencies, Health Tuskegee syphilis study
sciences Maintenance Plans, alternative healthcare practitioners,
patients, animal rights advocates, hospital administrations,
malpractice attorneys, news media
Social sciences Cultural or identity groups, privileged economic and social Brown v. Board of
classes, legislators, courts, admissions and ethics committees, Education (school
hate groups desegregation)
Industrial Industrial competitors, whistleblowers, labor unions, Tobacco industry cancer
research customers, consumer advocates, regulatory agencies, research
environmentalists
Weapons Enemy states, allied states, spies, terrorists, news media, social Manhattan Project
research activists
Table 2. Common epistemic adversaries
essential element of daily practice. These internal they (also) had experimentally confirmed the
dynamics may swing a project between the results announced by the visiting virologist at the
cooperative and adversarial modes. Therefore, conference, and this second group of biologists
during the course of scientific enterprises intended to publish independently right away.
researchers may face new or unexpected moral The first biology group communicated this turn
challenges. of events to the virology group, which began
Identifiable conditions make a biomedical writing immediately.
project swing either in the competitive direction In this narrow case we identify conditions
or the cooperative direction. Adversarial that support cooperative practices, such as
conditions, for example, include colleagues’ complementary skills, and conditions that
overlapping goals and/or methods; proximity to support adversarial practices, such as allocation
project completion and therefore to allocation of of credit at publication. The cooperation
credits; limited resources; and, at a personal between the first cellular biology team and the
level, researchers’ hidden agendas and breach of virology team was established due to
confidence. Cooperative conditions include (also) complementary expertise. The swing towards the
colleagues’ overlapping goals and/or methods of cooperative pole was enhanced by a longer stay
project; complementary skills or resources; the of one virologist with the cellular biology team.
need for face-to-face meetings; and, at a personal The desire to obtain credit made the project
level, friendship and trust. swing towards the adversarial pole and was
Our case study of competition and enlarged by a hidden agenda on the part of the
cooperation among three biomedical research virology team, who wished to publish alone first.
teams illustrates the natural fluctuation between Competition from another cellular biology team
the adversarial and cooperative poles. A virology made the project swing back towards the
research team sent a young researcher to study cooperative pole. Indeed, the methods and norms
abroad for a year with a cellular biology team, of science drive research projects along a typical
because of its expertise with a certain protein. trajectory of moral hazards and opportunities.
His hosts encouraged the visiting virologist to As this and other studies show (e.g., 6)
present his research results at a conference. They biomedical projects can be seen as intermediary
also urged him to publish the “hot” results between the adversarial and cooperative poles.
quickly, but he delayed. The biology team In particular situations, it can be very difficult to
subsequently confirmed the results distinguish adversarial from cooperative
experimentally and proposed to take a epistemic orientations. The model provides a
subordinate role in a joint publication with the point of reference by stating that the key
virologists. However, the virologists wished to difference between adversarial and cooperative
publish alone first. Meanwhile, a second cellular epistemologies is deliberate deception.
biology group contacted the first and implied that

301
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––

Utility of the Model for a project may cease to apply as the project
The epistemic model offers both conceptual and evolves. For an example from ethics of
practical gains to science ethics. Conceptually, application, the Manhattan Project authorized a
the model serves as a unifying schema for issues metabolic plutonium experiment on unwitting,
in science ethics. Two classes of scientific terminal patients, to gauge effects of plutonium
misconduct are commonly distinguished. The exposure on bomb production workers. In 1944
“ethics of research” is largely concerned with the many people would have agreed that the national
means of competition among researchers, such as security interest morally superseded the rights of
strategic secrecy. The “ethics of application” is patients, who were expected to die before the
concerned with the means used to attain scientific plutonium affected them adversely. But some of
and technological ends, such as creation of toxic the patients survived for decades and suffered
wastes (1, p. 503). These two classes are severe damages from plutonium injections,
distinguished by the types of harm produced. which invalidated the original moral rationale.
The epistemic continuum accommodates the For an example of the tracking problem from
ethics of research and the ethics of application in ethics of research, in our case study of three
a single schema. The harms change, but the biomedical research teams, the rationale for the
adversarial epistemic principles that lead to the project appeared to change during the course of
harms remain the same! Deception of colleagues the project. At first the advancement of
in recording data and deception of research knowledge was the ultimate objective, which
subjects in promising medical cures both follow includes the obligation to publish results as soon
the same adversarial epistemic principle of as possible. This objective was superseded in
deception of the adversary, although the later stages by the objective to obtain credit for
corruption of science and the injury to persons discovery. A key ethical requirement for a
are ontologically different harms. The epistemic scientific project would be to show how the
model identifies misconduct in science according original moral rationales, if needed, track along
to the principles of adversarial inquiry employed with the anticipated course of the project.
in the misconduct rather than the nature of the The fluctuation between cooperative and
harm. adversarial modes addresses the limitations of
Further, the model guides study of the front-end solutions to moral problems in science,
interaction between cooperative and adversarial such as voluntary informed consent of subjects
epistemic methods. Cooperative epistemic and authorship agreements. As a further
methods lead to specialization, perfection of contribution to science ethics codes, the
methods, and accountability in applications. epistemic model invites consideration of the most
Adversarial epistemic methods lead to expansion effective points of intervention for ethical codes.
of domains, innovation in methods, and speed of The model also suggests addressing potentially
application. To what extent are adversarial adversarial roles with support for the weaker
methods actually separable from cooperative party instead of only admonitions to the stronger.
methods in scientific projects? What are the For example, to moderate the potentially
costs and benefits of eliminating adversarial adversarial roles of researcher and graduate
methods? How can beneficial and destructive student assistant, ethical codes might standardize
competition be characterized? support for the student in the form of a mentor at
As a practical contribution to science ethics another institution.
codes, the model translates ethical problems in Philosopher Henry Sidgwick, who laid the
science—which philosophy of science cannot foundations for 20th Century ethics, considered
directly address—into products of a competing whether society would be more improved by
epistemology—which philosophy of science is correction of character flaws, so as to gain the
better equipped to address. For typical research capacity to follow our moral convictions, or by
projects, epistemic adversaries and collaborators moral understanding, so as to gain insight into
can be specified across the stages of the project, the consequences of our actions. Sidgwick (7)
and typical moral risks and opportunities can be advocated education of moral understanding on
assessed. the grounds that strong character coupled with
The model highlights what we call the conviction leads to the most serious moral
tracking problem: the original moral rationale offenses. Historically, this has been the danger

302
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Arrigo & Poulsen, Epistemic Model for Moral Hazards
for science. The epistemic model for scientific
misconduct follows Sidgwick in offering moral
understanding for science ethics education.

Acknowledgments
Arrigo acknowledges Harold William Rood for
consultations on military intelligence and Kurt
Smith for pointing out the symbiosis between
cooperative and adversarial epistemologies.
Poulsen was supported by the Danish Social
Science Research Council which she gratefully
acknowledges. Furthermore Poulsen extends her
thanks to the scientists who gave their time and
attention to her interviews. Jean Maria Arrigo is
largely responsible for the epistemic model.
Maj-Britt Juhl Poulsen is solely responsible for
the case study.

Bibliography
1. Forge J. Merton revisited.... . Metascience 1999;
8(3):503-504.
2. Poulsen MJ. Competition and cooperation: what roles in
scientific dynamics? International Journal of
Technology Management. In press. To appear in Vol.22,
No.8, 2001.
3. Arrigo JM. Sins and salvations in clandestine scientific
research: A social psychological and epistemological
inquiry. Claremont Graduate University; 1999.
Dissertation.
4. Magnier M. Japan Amazed as Archeologist’s Magic
Exposed as Sleight of Hand. Los Angeles Times 2000
November 9; Sect. A i 1, 13, 14.
5. Arrigo JM. The ethics of weapons research: A
framework for moral discourse between insiders and
outsiders. Journal of Power and Ethics [On-line journal]
1999; 1(4). Available WEB: http://spaef.com/JPE_PUB/
v1n4.html
6. Atkinson P, Batchelor C, Parsons E. Trajectories of
collaboration and competition in a medical discovery.
Science, Technology, & Human Values 1998; 23:259-
84.
7. Sidgwick H. The morality of strife. In: Practical Ethics:
A collection of addresses and essays. London: Swan
Sonnenschein; 1909. p. 83-112.

303
Scientific Misconduct as Organizational Deviance
Robert Dingwall, Institute for the Study of Genetics, Biorisks and Society, University of
Nottingham, UK

Keywords: Human subjects; Organizational deviance

Although, as Steneck points out in his background report for this meeting, scientific misconduct is
usually understood to involve “fabrication, falsification and plagiarism in proposing, conducting or
reporting the results of research”, human subjects protection cannot be excluded from this agenda.
There are two reasons for this. First, it may be argued that research misconduct is in itself a form of
human subjects abuse, since people have taken part in procedures that break the contract between
researcher and participants by not making a valid contribution to scientific knowledge. Second, as
Steneck also notes, integrity is a “measure of the degree to which researchers adhere to the rules or
laws, regulations, guidelines and commonly accepted professional codes and norms of their
respective research areas.” To the extent that human subjects protection is the objective of much of
this regulatory framework, we may argue both that researchers who compromise on the truthfulness
of their reporting may be more likely to commit other abuses and that the success or failure of
strategies for human subjects protection may offer relevant lessons for strategies to limit misconduct.
The death of Jesse Gelsinger in the course of a gene therapy trial at the University of
Pennsylvania Institute for Human Gene Therapy (IHGT) in September 1999 has cast a long shadow
over the adequacy of the regulatory framework in this area of medical science. It has led to
significant restructuring of IHGT, has been used to justify changes in Federal regulatory structures
and has provoked a bout of intense internal and external scrutiny of practice in clinical trials
throughout the international community. While the narrative of events at IHGT is now reasonably
well-established, there is still much to be understood about the reasons for the regulatory breaches
brought to light by the subsequent investigations, particularly given the lack of evidence for any
causal relationship between these and Gelsinger’s death. How significant are the breaches identified?
If they are relatively insignificant, have the correct regulatory conclusions been drawn? Will the
changes proposed or introduced through the spring and summer of 2000 actually make trials safer, as
opposed to satisfying public and political demands that “something be done?”
Traditionally, failures of the kind represented by the Gelsinger case have led to a search for
blameworthy individuals, whose errors or omissions produced the negative consequences that have
given rise to public scandal. The conventional response has been to call for individual sanctions and
a strengthening of regulations or their enforcement. However, social scientists have become
increasingly critical of this approach, arguing that organizational failures or misconduct are nowadays
rarely the result of individual negligence or deceit. More typically, these failures arise as the
unintended consequences of personnel carrying out their routine work under conditions of

Corresponding author: Robert Dingwall, Institute for the Study of Genetics, Biorisks and Society, Law and Social Science
Building, University of Nottingham, University Park, Nottingham NG7 2RD, UK, +44-0115 951-5418 (voice), +44-0115
846-6349 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
organizational or environmental complexity that public concern about the possible implications of
fail to give them appropriate feedback on the the release of genetically engineered material. In
implications or results. Policy responses that the US, National Institutes of Health (NIH)
increase complexity may actually further obstruct established the Recombinant Advisory
feedback, or introduce new opportunities for Committee (RAC) to oversee development.
unpredictable system interactions to occur, rather However, RAC’s formal powers were limited,
than eliminating those that proved troublesome in and unlicensed experimentation took place as
the past. This argument, originating with the long ago as 1980, although the clinician involved
work of Charles Perrow (1) in the US and Barry was heavily censured. The first FDA approved
Turner (2, 3) in the UK, has been developed over trial began in September 1990, to treat an
recent years by Diane Vaughan (4, 5) in her inherited immune disorder, and more than 400
studies of the 1977 Ohio Revco Medicaid fraud trials are known to have taken place, worldwide,
and the Challenger space shuttle disaster. In the during that decade. However, clinical benefit has
latter, for example, Vaughan shows how the been hard to demonstrate. In 1995, Harold
social structure of NASA and its contractors, and Varmus, then Director of NIH, created an ad hoc
the dispersion of information about problems committee to review NIH investment in a field
with the O ring seal, allowed correct engineering that seemed to have so much potential and to be
reasoning to produce advice to launch that had realizing so little of it. This committee reviewed
devastating consequences. For present purposes, more than 100 approved protocols but its report
however, the Revco study may be a more useful to the RAC meeting in December 1995
model with its deliberate attempt to merge the underlined the lack of progress and the
understandings of social scientists who have fundamental scientific problems that remained
studied organizations, regulatory bodies, and unsolved.
white collar crime. How do “respectable folks” Coincidentally, the IHGT trial was approved
end up in situations where they breach at the same RAC meeting. The trail was intended
regulations intended to keep them honest? Why to investigate possible treatment for a condition
do organizations fail to prevent this? known as ornithine transcarboxylase deficiency
This paper falls into three parts. The first (OTCD). This condition arises when a baby
briefly reconstructs the Gelsinger case from inherits a broken gene that is needed for the liver
published sources available over the Internet. (It to produce an enzyme that breaks down
is not claimed that this is an exhaustive account, ammonia. The IHGT researchers wanted to
given the time and resources available.) Some of package this gene with a modified adenovirus
the main ideas put forward by Vaughan are then and inject it into the hepatic artery to get the most
introduced, as a way of thinking about the kind direct delivery to the liver. Although there were
of issues represented by this incident. Finally, some anxieties expressed about this delivery
these ideas are used to look at the Gelsinger route, both RAC and FDA eventually agreed to
narrative, with some reference to a brief period of approve the trial. In 1999, Jesse Gelsinger was
participant observation in a British university’s the eighteenth and final patient to be recruited.
genetic science laboratories during summer 2000. Gelsinger was eighteen years old and in good
health at the time but could not be described as a
Gene Therapy at the IHGT healthy teenager. He had a long history of
According to an official Food and Drug OTCD problems, which had finally been brought
Administration (FDA) version (6), although gene under some control by a combination of
therapy is an attractive idea, it has been slow to medications and a highly restricted diet. He
fulfil its theoretical promise. It has proved received the experimental treatment in September
difficult to package correctly-functioning 1999 and died four days later, apparently from an
versions of disease-related genes in a way that overwhelming immune response to the carrier
allows them both to be delivered into the virus.
appropriate cells of a patient and to switch on. The subsequent FDA investigation found a
US researchers have generally looked to series of regulatory breaches committed by the
modified adenoviruses as the delivery vehicles, IHGT (7). Gelsinger had been entered into the
although UK researchers have been more trial as a substitute for another volunteer,
attracted by lipids. The general principles have although his high ammonia levels at the time of
been known since the 1970’s, giving rise to treatment should have led to his exclusion. IHGT
306
––––––––––––––––––––––––––––––––––––––––––––––––––––– Dingwall, Scientific Misconduct as Organizational Deviance
had failed to report serious side effects was reconstituted as the Office for Human
experienced by two previous patients in the trial, Research Protections (OHRP), as advised by an
and the deaths of two monkeys given similar NIH review submitted in 1999 before the
treatment had not been mentioned to Gelsinger or Gelsinger incident. At the same time, the newly
his father at the time informed consent was constituted OHRP was given expanded authority
obtained. FDA shut down the OTCD trial and relocated in the Office of the Assistant
immediately. FDA Form 483 issued to Dr. James Secretary for Health in the Department of Health
Wilson, IHGT Director, on January 19, 2000, and Human Services (DHHS), placing it closer
listed a number of concerns, which were to the line of direct political authority. The
summarized in a letter from FDA dated overall response was summarized in evidence to
January 21, 2000, as failing to ensure the a US Senate Subcommittee on May 25, 2000,
following: under five headings: education and training;
conduct of the study in accordance with the informed consent; improved monitoring;
clinical protocols that are contained in the IND; conflicts of interest; and civil money penalties.
obtaining adequate informed consent from All clinical investigators receiving NIH funds
subjects prior to participation in a study of an would have to show that they had received
investigational agent or performance of appropriate training in research bioethics and
invasive procedures; compliance with reporting
protocol changes and adverse events to the
human subjects protection, as would Institutional
responsible IRB; filing of safety reports as Review Board (IRB) members in their
outlined in 21 CFR 312.64; and maintenance institutions. Audits of informed consent records
of complete and accurate records (8). would be performed and IRBs would be required
This letter suspended authorization for all IHGT to monitor informed consent elicitation more
clinical trials. A nationwide review of other closely. Informed consent would have to be re-
approved trials revealed a high level of under- confirmed after any significant trial event. A
reporting of serious adverse events and possibly wider range of Clinical Trial Monitoring Plans
associated deaths. General shortcomings would have to be reviewed by both NIH and
included: eroded adherence to requirements or local IRBs. Conflict of interest rules for
standards of informed consent; lack of investigators would be reviewed to ensure that
investigator adherence to good clinical practices research subjects and findings were not
and current Federal requirements; lack of manipulated for commercial gain. Finally, as
adequate quality control and quality assurance mentioned earlier, legislation would be proposed
programs for the gene therapy products used in to allow FDA to levy civil fines for regulatory
trials; weak IRB processes; financial conflicts of breaches (9,10).
interest; lack of public access to safety and Meanwhile, IHGT and the University of
efficacy data; limited regulatory enforcement Pennsylvania had initiated their own actions.
options for Federal authorities; inadequate IHGT filed a response to FDA Form 483 on
resources for enforcement; scope for improved February 14, 2000. In contrast to the FDA
co-ordination between FDA, NIH and OPRR; version, IHGT noted that it had promptly
and poor understanding by investigators of FDA informed FDA, RAC, and the relevant IRB of
and NIH roles in gene therapy oversight. Several Jesse Gelsinger’s condition and that, in contrast
other trials were suspended for regulatory to the FDA version above, IHGT had taken the
breaches or because of technical similarities to initiative in suspending the trial. Moreover,
the OTCD trial. Other funders also suspended IHGT could demonstrate that every trial
trials for review (9). participant had given informed consent and their
In March 2000, FDA and NIH launched a eligibility for participation was fully
Gene Therapy Trial Monitoring Plan, increasing documented. There had been delays of 3-4
reporting requirements and requiring researchers months in submitting toxicity information on
to communicate more with each other about some early participants, which should have been
safety issues. In May 2000, President Clinton discussed with FDA before proceeding with the
announced plans for legislation to allow FDA to next cohort. Nevertheless, FDA had these
impose civil penalties on researchers and reports in its possession for more than six months
institutions for regulatory violations. In June prior to August 1999 when it approved the trial’s
2000, the NIH Office for Protection from continuation for the cohort that included Jesse
Research Risks (OPRR), established in 1972, Gelsinger. IHGT had Standard Operating
307
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Procedures that met the regulatory requirements the ORA would monitor the trials themselves or
in force. The study in which two primates had recruit external consultants to do so. The IHGT
died was unrelated, using different genetic vision of a combined unit for basic, pre-clinical,
material to treat a different disease. One primate and clinical work in gene therapy would be
had shown a mild reaction to a viral vector from abandoned. The Center for Bioethics would
the same generation but at a much higher dose– become a free-standing department. IRB
seventeen times higher–than in the OTCD trial. procedures would be strengthened and given
Available evidence did not establish any causal extra resources. Ultimately, principal
link between Gelsinger’s plasma ammonia level investigators and research coordinators would
prior to the infusion and his death (11). FDA require certification before being allowed even to
reacted critically to the IHGT response. In a submit protocols to the IRB. The University
Warning Letter on March 3, 2000, there was a already restricted investigators from having
parallel exchange over the non-clinical financial stakes in companies sponsoring trials
laboratories at IHGT (12). but would review and strengthen this restriction.
The University President set up an At the time of writing (October 2000), a
independent external panel to review IHGT. The number of loose ends remained, particularly the
panel reported on April 27, 2000 (13). The panel final determination of FDA’s response to IHGT
noted the discrepancies between the FDA Form and University of Pennsylvania’s actions and the
483 and the IHGT response but disclaimed nature of any new legislation. However, there is
sufficient regulatory expertise to comment. The no doubt that the Gelsinger case has come to be
panel focused on the operations of IHGT, noting seen as iconic of problems in the regulation of
its commitment to good practice and any scientific research and of public and political
necessary revision of operating procedures. mistrust of this process, not just in the US but
IHGT had already contracted out the monitoring also in the UK and other countries with advanced
of its trials to an independent organization. levels of science. The regulatory and
However, the panel noted the growing costs of institutional responses will be widely studied.
compliance and the need for the university to How much faith should we place in them?
invest more resources in this area. The panel
made the following recommendations. The Understanding Organizational
university needed better internal monitoring and Misconduct
lower workloads for each of its IRBs. Over the last thirty years, researchers in the fields
Bioethicists should cease to be involved in of law and society and of organizational studies
operational decision-making but act as have become increasingly sceptical about the
consultants to investigators who would be effectiveness of regulatory interventions as
responsible for their own actions. Conflict of incentives for corporate bodies to act in a lawful
interest policies should be reviewed. There fashion. Vaughan has summed up the alternative
should be closer scrutiny of informed consent as a view that organizational misconduct is
procedures to ensure compliance with the letter produced by social structure:
as well as the spirit of FDA regulations. The By social structure, I mean (1) the stable
panel also questioned the lack of continuing characteristics in American society that form
review for university institutes, the wisdom of the environment in which organizations
concentrating all gene therapy work in one conduct their business activities: sets of social
organization, the training of young clinical relations, laws, norms, groups, institutions; and
investigators in the special issues of (2) the stable characteristics of organizations
investigational drugs, and the desirability of the themselves: internal structure, processes, and
the nature of transactions. (4, p. 54)
university itself being simultaneously involved in
Vaughan elaborates on a model first suggested by
the production of vectors, research, and the
Merton (15) that locates the incentives for
monitoring of standards. The President’s
deviant action in the tension between cultural
response was delivered on May 24, 2000 (14).
goals of economic success and social structures
She announced a new assessment of all clinical
that limit access to legitimate means for
trials by the University’s Office of Regulatory
achieving these goals. Merton set out a range of
Affairs (ORA). Where regulatory affairs
possible responses, but the one that interests
professionals were not already involved, as in
Vaughan is “innovation”. This is the attempt to
trials sponsored by pharmaceutical companies,
achieve the valued goals by expedient but
308
––––––––––––––––––––––––––––––––––––––––––––––––––––– Dingwall, Scientific Misconduct as Organizational Deviance
prohibited means, justified on the basis that the visibility of their actions. Complex organizations
unequal access to legitimate means compromises multiply opportunities for misconduct through
the norms that distinguish legitimacy from their structural differentiation and task
illegitimacy. If this distinction is perceived to be segregation.
arbitrary or discriminatory, then it may fail to The result is what Vaughan terms “authority
command moral respect. In the context of leakage”, the loss of capacity for internal control.
science, for example, Barber and colleagues (16) The actions of subunits may become effectively
showed that those most likely to cheat on the invisible, particularly where they involve
norms of the professional community were those specialized knowledge that is not shared
who felt unjustly treated in their careers. elsewhere in the organization. A rational process
Vaughan notes that Merton focused mainly on of internal censorship designed to match upward
the impact of the tension between culturally information flows to the processing capacity of
valued goals and social structures for individuals senior managers, obscures misconduct, and
in lower social classes. However, Vaughan argues diffuses personal responsibility. Finally, the
that this approach is at least as well suited to the nature of transactions both provides legitimate
analysis of organizations, which may be more opportunities for illegitimate behavior, and
strongly driven than individuals by the further minimizes the risk of detection and
requirements of profit-maximization but where sanctioning. Transactions between complex
competition undercuts the force of norms. The organizations have four distinguishing
processes of change that are the dynamic of a characteristics: formalization; complex
market economy continually challenge the processing and recording methods; reliance on
normative order of that economy. The trust; and general rather than specific monitoring
formalization of norms into law has limited procedures. Because of the difficulty of
effectiveness. Legal responses to “innovation” monitoring each individual transaction,
occur after the event and are skewed by the organizations tend to rely on signals that can be
extent to which both rules and their enforcement manipulated to present an appearance of
rest on negotiations between regulatory agencies legitimacy to outside observers, whether
and the firms they regulate (17). transaction partners or regulators.
As Vaughan points out, unlawful behavior Vaughan discusses the particular example of
cannot be explained solely in terms of these Medicaid fraud where determinations of
social structural tensions. Opportunities must eligibility for participation tend to rest on data
arise that offer the possibility of unlawful acts submitted by would-be service providers. The
and the regulatory environment must be such that complexity of the government paperwork and the
there is a reasonable chance of escaping lack of resources for verification create
sanctions. Vaughan points to the processes, conditions where willful misrepresentation can
structures, and transactions of modern complex occur. This also indicates a problem of system
organizations as the sources of opportunity. As interface, where the culture and structure of two
the literature on white-collar crime shows, these organizations, in this case government
create the conditions for individuals to act bureaucracies and relatively small for-profit
illegitimately: her claim is that they also make enterprises, conflict. If these cannot be brought
organizational misconduct possible. into alignment, one or both organizations may
Organizational processes create a moral and choose unlawful actions as a means of achieving
intellectual world for members, encouraging their goals. Vaughan notes how Revco
them to identify with the organization and its executives felt justified in false billing the Ohio
goals. The survival of one becomes linked to the Welfare Department for an amount equal to the
survival of the other. Those most exposed to claims for payment that had been denied on what
temptation are those in the subunits most relevant Revco felt to be excessively bureaucratic
to the resource or profit-seeking goals, with grounds. The Welfare Department wanted Revco
information linking subunit performance to the to internalize a government agency culture that
achievement of those goals and some Revco found incompatible with a private, for-
responsibility for that achievement. Their choices profit enterprise.
reflect their awareness of the organization’s
relative rewards for achievement and its Regulating Science
sanctions for illegality and of the structural Vaughan’s analysis of the Revco case focuses on
309
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
the potential sources of misconduct in profit- managing risk. It can work reasonably
seeking organizations, although she makes some effectively in mature environments where risks
suggestions about its possible relevance to other are well-understood. In many engineering
sorts of enterprise. Scientific research situations, for example, there is a recognizable
organizations have some peculiar features, and cycle of risk and regulation. A new technology
may vary somewhat according to whether they generates a number of accidents that lead to a
are in universities, not-for-profit corporations, or definition of hazards and a regulatory response
commercial companies. However, it is arguable that produces a safe environment until the next
that, whether or not scientists are overtly engaged significant change in technology comes along.
in profit-seeking, the incentives that they face are Although there are also routines in scientific
functionally equivalent. Profit, as Vaughan notes, research, science is ultimately about pushing into
is merely the most obvious indicator of an the unknown and taking unknowable risks. A
organization’s success in locating and securing regulatory regime that prevented all risk would
resources for its operations and survival. prevent all scientific innovation. However, to the
Scientific work depends upon flows of grant and extent that contemporary societies have a low
contract income which, in turn, depend upon the tolerance for risk, there is an inherent tension for
production of results which lead to further regulators between the demand that risk be
income flows. These may derive from patentable averted and the functioning of the regulated
innovations or from peer esteem, which leads to enterprise at all. A level of regulation that stifles
publication in high-quality journals, professional enterprise is not in the regulators’ interest any
networking opportunities and so on. For the more than a failure to regulate sufficiently that
individual scientist, personal rewards may be leads to legitimacy problems with the public or
symbolic rather than material, but these virtual the political system. In any clinical trial,
profits are converted into economic resources for participants assume some measure of risk:
the research organization (18). Science is regulators may do their best to manage this, but it
reward-driven in the same way as other cannot be eliminated because of the variability of
enterprises and, as elsewhere, a failure to win human response and possible interactions with
rewards leads to bankruptcy, whether personal or other idiosyncratic features of the participant’s
corporate. In the British university department biology or environment. The question is whether
that I studied, for example, laboratories began participants are adequately informed about this
almost literally as shells, which faculty were risk and compensated for adverse outcomes. If
expected to equip for both the capital and the risks were eliminated, so would be the
consumable needs of their research through their possibility of discovery. Regulators must always
income-generating activities. A run of trail behind and the letter of regulation can never
unsuccessful grant applications could lead to a be more than a partial solution to the
downward spiral where the investigator simply management of risk.
ran out of resources. The department claimed to If the effectiveness of regulation is
be unusual in having an internal taxation system necessarily limited, we may need to look more
that could provide some support for a member in closely at the social norms of research
this position, at least for a period, in the hope that organizations and the structures in which they are
their luck would turn. This was said to be embedded (19). The university department that I
unpopular with funders who would have studied was a relatively compact physical group,
preferred to see a purer market system with no where the principal investigators had offices in
socialization of resources. the corner of the laboratories in which their
If this leads us to accept that Vaughan’s postdocs, research assistants, technicians, and
analysis could be broadly applicable, we also graduate students worked. Laboratory work was
need to acknowledge that there may be some highly visible to colleagues. There was also an
differences between scientific research active tradition of seminars, journal clubs,
organizations and other kinds of enterprise. The gathering for coffee and lunch breaks, and
most important may be the way in which the departmentally-based socializing. This facilitated
problems of the reactive nature of regulation are the development of a departmental culture,
accentuated by the defining characteristic of although it did not prevent perceptible
science, namely its engagement with uncertainty. differences emerging in the climate of different
Regulation is an institutionalized means of faculty member’s laboratories. Clinical trials,
310
––––––––––––––––––––––––––––––––––––––––––––––––––––– Dingwall, Scientific Misconduct as Organizational Deviance
however, as the Gelsinger documents clearly that, although not recognized by the FDA,
show, tend to have a much longer chain of adequate documentation for consent does exist.
command, which makes important parts of the However, the elicitation of consent is also a
process substantially invisible to principal difficult interactional task. How do you ask
investigators. someone voluntarily to assume a risk that can be
The scale and complexity of the clinical trial broadly described but is ultimately unknowable
process has generated an increasingly intricate until after the event. Lower-level personnel
division of labor. At the top are the principal charged with the execution of the task tend to
investigators (PIs), whose strategic vision and deal with this by a measure of improvisation.
social networks are crucial to generating the flow They seek to comply with the spirit of the
of resources that keep the enterprise going. In regulation rather than the letter.
the middle are the trial managers and The result is a degree of variance that is hard
coordinators who keep the process on track. to reconcile with the command and control
Patients, however, actually have direct contact approach. Both the University of Pennsylvania
with much lower level people who obtain and FDA seem to have responded by trying to
informed consent, administer the interventions, toughen the regime. Indeed there are even
and collect the test data on the results. The proposals that IRB members should monitor the
“hired hand” problem has long been recognized consent process by direct observation. The
by those social sciences that make extensive use problem would seem to be that you could reduce
of survey techniques (20). How do you guarantee the process to a script, force the consent-takers to
that low-level workers doing rather mundane read the script aloud to the patient by recording
jobs do not simply make up data or ignore the or observing them, as in call centers, and then
code book when entering it? Computerized discover either that hardly anyone is willing to
interview techniques have reduced the volunteer, because the process has been made
opportunities for misconduct, but it has regulator-friendly rather than user friendly, or
historically been a considerable challenge to the that consent is formal rather than substantive and
management processes of survey organizations. that patients who experience adverse outcomes
It represents the same problem of authority can still reasonably claim to have been deceived
leakage and internal censorship that Vaughan or not to have understood the nature, purpose,
describes. Structural differentiation and task and risk/benefit ratio of the trial.
segregation make operational performance In effect, this reproduces the Revco problems
invisible to senior managers. Whatever of the organizational interface between a Federal
performance or quality standards are set, regulatory bureaucracy and, in this case, the
managers are unable to follow them through. At professional traditions of university science.
the same time, information from lower-level Traditionally, universities have been federations,
personnel is censored as it rises to match the or even confederations, of professionals, with a
capacity of supervisors and managers to handle high degree of internal autonomy and limited
it. collective responsibility. Although this model
Various solutions have been tried, two of has come under some pressures from demands
which are worth further discussion here. One is for greater social accountability in recent years,
more detailed organizational rule-making to try these have been opposed by the encouragement
to govern lower-level personnel by command and of entrepreneurial science. The difficulties of
control methods. The result of this is usually to raising student fee income to a level where
reduce further commitment to organizational salaries competitive with the general
goals and to sacrifice the potential gains from a commercialization of professions (21-23) can be
degree of flexibility at the point of operational paid have been met by a shift in culture that
activity. If we take the specific example of allows those who can to top up their incomes
informed consent, this has become the subject of with consultancy earnings and stakes in spin-off
increasingly elaborated procedural rules. companies. Although academics may be able to
Consent may now be deemed to be informed raise their market price by trading on their
only if it is in accordance with these rules, university’s reputation, they are simultaneously
something that may account for the discrepancy less constrained by the university’s employment
in view between FDA and IHGT. FDA finds that discipline, since their salary may be a relatively
the paperwork is not in order, while IHGT claims small proportion of their income. This poses a
311
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
considerable management problem for look behind the letter of such contracts to the
universities, since bureaucratization may cost responsibility of those issuing them to audit the
them faculty whose presence is crucial to their performance of contractors. The growing
general competitive position. The University of liability of hospitals for the acts of physicians
Pennsylvania, for example, proposes to introduce afforded admitting privileges is an obvious
certification for PIs: if this is perceived as parallel. The result is likely to be an
burdensome, the result may be that the university organizational internalization of law, as the
loses star talent to less intrusive competitors. alternative to bureaucratization, with PIs required
The result, as is evident from the FDA to attend to the compliance of the documentation
response to the Gelsinger events, is often a of their work with the forms of private rather
division of rules into those taken seriously and than public law (24). It is simply a different kind
those on the book but disregarded unless of interface problem.
something goes wrong and a source of Ultimately, there is probably no substitute for
sanctioning is required. There is a hierarchy of the more active engagement of PIs with their
rules, some of which “really” matter and some of projects and methods of countering authority
which are there for use only if needed. The leakage and internal censorship. The paradox is
IHGT/FDA clashes seem to suggest that that the enhanced systems of scrutiny, whether
something similar has happened. Having bureaucratic or legal, will tend to make this more
complied with what IHGT seems to have been difficult by enhancing the competing calls on this
led to understand were the “important” rules, it pool of senior investigators to participate in peer
clearly feels aggrieved that the FDA inspection oversight of others. To the extent that their time
has produced an exhaustive list of breaches, is drawn into this system, by the sorts of
arguably to cover the agency’s own collusion in measures that FDA envisions in terms of more
the procedures at the Institute. One might note frequent sharing of trial experiences or the
particularly the counter-charge that FDA had expansion of IRB membership to spread
been in possession of toxicity reports on earlier workload and allow more intensive scrutiny of
trial participants for six months without comment proposals, then the problem that internal
before approving the recruitment for the final censorship solves will grow worse. Internal
cohort that included Gelsinger. censorship, remember, is the solution to the
When bureaucratic command-and-control limited time and attention that senior
fails to defend the organization from regulatory organizational actors can give to any particular
pressures or liability suits, one response can be problem. If time becomes more restricted, then
its replacement by a network of outsourced sub- censorship will increase. The FDA’s measures
contractors, as the University of Pennsylvania may mean that PIs become much better informed
seems to envisage. PIs or research organizations about other people’s problems and less well
lay off the risk by sub-contracting the work informed about their own. Which is most likely
through contracts that specify performance and to contribute to safer research for human
quality but locate the responsibility outside the subjects?
core business. The difficulty with this model is This is obviously a brief account of a
that exhaustive performance contracts are complex story that is still some way from
essentially impossible to write and that further completion. It is also heavily reliant on the
incentives for misconduct tend to be created. If a public record and would obviously benefit from
sub-contractor is required to deliver a certain interview data of the kind that Vaughan had
number of patients and associated paperwork for access to in her work. However, it may serve to
a fixed price, they clearly have reason to see exemplify an approach to the study of scientific
where corners can be cut. The PI sacrifices misconduct and, in particular, to illustrate some
control over data quality and, to some extent, of the very real difficulties of imposing a strong
ethics in favor of protection from the professional external regulatory regime on practice. The
or legal implications of failing to control either issues of compliance that arose in the human
personally, provided that there are adequate risk- subjects protection of Jesse Gelsinger are
shifting clauses in the original contract. It is, immediately parallel to those that arise in
however, probably naive to assume that such controlling falsification, fabrication, and
risk-shifting will be an effective defense, plagiarism, which also are compromised by the
particularly given the tendency of US courts to structural and cultural problems that lead to
312
––––––––––––––––––––––––––––––––––––––––––––––––––––– Dingwall, Scientific Misconduct as Organizational Deviance
13. Report of the Independent Panel Reviewing the
authority leakage and internal censorship. It is
University of Pennsylvania’s Institute for Human Gene
only by recognizing and engaging with these Therapy. University of Pennsylvania Almanac 2000
underlying problems that effective interventions May 30; 46 (34).
can be designed. 14. Action by the University of Pennsylvania in Response to
“Report of the Independent Panel Reviewing the
Acknowledgements Institute for Human Gene Therapy”. University of
Material on the Gelsinger case was assembled by Pennsylvania Almanac 2000 May 30; 46 (34).
Mark Allan. I am grateful to Diane Vaughan and 15. Merton RK. Social Structure and Anomie. In: Merton
RK, editor. Social Theory and Social Structure. New
Liz Sockett for their comments on an earlier
York: Free Press; 1968. p.185-214.
draft. 16. Barber B, Lally J, Makarushka JL, Sullivan D. Research
on human subjects: problems of social control in
Bibliography medical experimentation. New Brunswick, NJ:
1. Perrow C. Normal Accidents: Living with high risk Transaction Books; 1979.
technologies. New York: Basic Books; 1984. 17. Albiston C. The rule of law and the litigation process:
2. Turner BA. Man-made disasters. London: Wykeham the paradox of losing by winning. Law and Society
Publications; 1978. Review 1999; 33 (4): 869-910.
3. Turner BA, Pidgeon NF. Man-made disasters. Second 18. Latour B, Woolgar S. Laboratory life: the social
ed. Boston: Butterworth-Heinemann; 1997. construction of scientific facts. London: Sage; 1979.
4. Vaughan D. Controlling unlawful organizational 19. Etzioni A. Social norms: internalization, persuasion and
behavior: social structure and corporate misconduct. history. Law and Society Review 2000; 34 (1): 157-178.
Chicago: The University of Chicago; 1983. 20. Roth JA. Hired hand research. American Sociologist
5. Vaughan D. The challenger launch decision: risky 1966; 1: 190-196.
technology, culture and deviance at NASA. London: 21. Hanlon G. Professionalism as enterprise: service class
The University of Chicago Press; 1996. politics and the redefinition of professionalism.
6. Human Gene Therapy > Harsh Lessons, High Hopes. Sociology 1998; 32 (1): 43-63.
FDA Consumer Magazine 2000 Sep-Oct. 22. Hanlon G. The changing nature of professionalism and
7. Statement of Jay P Siegel, MD, Director, Office of the fracturing of the service class. International Review
Therapeutics Research and Review, Center for Biologics of Sociology 1999; 9 (1): 87-100.
Evaluation and Research, Food and Drug 23. Hanlon G. The commercialisation of accountancy.
Administration, Department of Health and Human Basingstoke: Macmillan, 1994.
Services before the Subcommittee on Public Health, 24. Edelman LB, Suchman MC. When the ‘haves’ hold
Committee on Health, Education, Labor and Pensions, court: speculations on the organizational internalization
United States Senate, February 2, 2000. of law. Law and Society Review 1999; 33 (4): 941-991.
8. Letter from the FDA to the IHGT. University of
Pennsylvania Almanac 2000 Jan 21; 46 (18).
9. Statement by William F Raub PhD, Deputy Assistant
Secretary for Science Policy, US Department of Health
and Human Services before the Subcommittee on Public
Health, Committee on Health, Education, Labor and
Pensions, United States Senate, May 25, 2000.
10. Statement of Greg Koski, PhD, MD, Director, Office for
Human Research Protections, Office of the Secretary,
Department of Health and Human Services for the
Hearing on Human Subjects Protections in VA Medical
Research before the Subcommittee on Oversight and
Investigations, Committee on Veterans Affairs, US
House of Representatives, September 28, 2000.
11. Summary of IHGT response to the FDA. University of
Pennsylvania Almanac 2000 Feb 22; 46 (22).
12. Warning Letter from Steven A Masiello to James M
Wilson, March 3, 2000. Warning Letter from Steven A
Masiello to James M Wilson, July 3, 2000. [I have been
unable to locate the university’s response to the first of
these or details of the Form 483 issued on or about 1
March 2000 and the university’s response on May 4,
2000, which precede the second letter.]

313
A Market Approach to Research Integrity
Aditi Gowri, University of Texas, Austin, USA

Keywords: Buyer’s market, Collective agency, Ethics, Market demand , Research integrity, Research
funding

Introduction: Supply Side Ethics

The standard approach to research integrity can be characterized as supply side ethics because of its
emphasis on individual persons who supply research products. These suppliers—researchers—are
screened, educated, exhorted, given incentives, and as a last resort threatened in the interest of getting
them to be honest, to keep careful, accurate records, to test their hypotheses fairly, to use reporting
techniques fairly, and to train their subordinates and encourage their colleagues in turn to do likewise.
This paper discusses the effects of the demand for research on the collective integrity of research
fields.
Briefly, the argument is as follows: where research takes place within a market for research
products, effective demand in this market will affect the distribution of knowledge produced.
Therefore, the body of scientific knowledge will be skewed by demand. The analysis here suggests
that this skew can result in a form of malignant bias resulting from demand. Paradoxically, this form
of bias occurs in the absence of corrupt researchers, research designs or grantors. It is a form of bias
nonetheless, since it can lead to misleading research-based knowledge and less than optimal policy
decisions. Thus, it should be of concern to researchers in research ethics.

There is a Market for Research Products

In the market for research, scientists are the suppliers and those who offer to pay research expenses
through grants or contractual funding are the “consumers” of research products. Any offer to fund
research activities constitutes a demand for the kind of research that is expected to result from those
activities. Principal Investigators can be seen as entrepreneurs who compete with their peers for
contracts. Those who are successful in obtaining contracts become suppliers and then use the
resources they have obtained to hire labour and buy raw materials—essential components in the
production of research.

Objection: funding cannot produce science

It is normal to accept that there is a competitive market for research funding but to separate this
conceptually from the idea of a market for research products. Knowledge, as usually understood, is
not something that can be sold to order like a car or a bunch of carrots. On this standard model,
scientists choose to pursue particular research questions because of their intrinsic value and expected
fertility. Funding comes to those with skill who choose a fruitful line of inquiry, as a reward for past

Corresponding author: Aditi Gowri, LBJ School of Public Affairs, University of Texas, Austin, TX 78713-8925, 512-471-
8245 (voice), 512-471-1835 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
successes and to support the promise of future to doing research over the course of a
productivity. But there is no simple buyer-seller professional life. On the other hand, a senior
transaction. Any appraisal of research products researcher with better funding is likely to attract
themselves must be based on examination of the more and brighter young scholars than his less
autonomous technical pursuit of the research generously funded colleagues. Thus the demand
craft: experimental design, data collection, record for knowledge, operating through the demand for
keeping and the interpretation of results. There is junior collaborators and research assistants, plays
of course an acknowledgement that getting a part in developing the competences and
research funding is a competitive pursuit, but the commitments of each new generation of
funding transaction is seen as completely researchers. Demand not only has an immediate
external to the generation of research products. market effect but also a life cycle effect on the
In other words, there may be a market for researcher’s capacity for—and commitment to—
funding but there is no market for knowledge. future research projects. Again, it is hardly
This conventional separation is unsatisfactory blameworthy for a junior researcher to consider
because funding affects the actual content of the size of available fellowships before choosing
research products in at least three ways. to work in a particular sub-discipline or
laboratory.
How does demand affect research
knowledge? Demand calls forth its supply
First, some researchers will modify their research In all of these ways, the economic demand for
questions, design and methodology to receive research will affect the supply of research
funding. It is not difficult to think of colleagues products developed. To accept this conclusion
who have changed their research questions or we do not have to believe that the demand for
design slightly to obtain the interest of a funding research can produce its own supply (although
agency. Indeed, at least in the social sciences, this is the way an economist might put it), nor
often the tail wags the dog, with research that research in the absence of funding is
proposals and even research programs developed impossible. We must only accept that some
in response to offers to fund. Generally this is researchers will respond to the incentives offered
not considered to be dishonorable, provided the by granting agencies and that those who do so
proposed studies are intrinsically legitimate and will be better situated to generate research than
carried out fairly. the rest. In other words, research flourishes in
Second, researchers who really want to the presence of money, and generating research
pursue research interests or designs that do not fit products without money is very difficult and rare.
some effective demand for knowledge are like Grantor sovereignty certainly is not absolute; it is
sellers in a market with no buyers. They can still no more than a form of consumer sovereignty,
conduct research, but only to the extent that resulting from the prerogative of buyers in any
personal funds or their general institutional market to demand the products that give them
budgets are adequate to support its costs. In most satisfaction.
general, smaller budgets will limit the scale and
type of work they can do. Since there is always Contrast the demand for corrupt research
competition for scarce research dollars, grantors It is worth emphasizing that the effect of demand
have the prerogative of declining proposals from on knowledge does not entail any individually
researchers who do not offer what is demanded. discreditable conduct on the part of either buyer
In other words, the market for research is a or supplier. A demand for corrupt research
buyer’s market. products probably exists. For instance, a grantor
Third, if we look at a researcher’s life cycle, with a preferred ideology may put pressure on a
the effect of demand is most strongly felt at the researcher to design not quite fair tests of
early stages of a career. Doctoral and post- hypotheses, to address data selectively, or to
doctoral researchers usually must serve as misreport or over-generalize findings. Perhaps
apprentices to a more senior researcher to begin more subtly but no less deceitfully, a
to earn a living in the research trades. Because it pharmaceutical company might commission
is an apprenticeship phase, junior researchers are more than one study of a drug, publicize only
expected to develop a package of skills and those favorable to its product, and bury the rest.
competences that will then affect their approach Each of these is an example of corrupt(ing)
316
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gowri, A Market Approach to Research Integrity
demand, but neither is our concern here. While purchaser, they also become part of the common
the demand for corrupt research is certainly stock of knowledge. Research produced for one
worthy of study, a discussion of its extent and purpose will often have unexpected “external”
effects does not lie within the scope of this paper. benefits and uses. (Proprietary approaches to
Throughout this discussion our concern is rather knowledge present only an apparent challenge to
with the demand for legitimate, honest research this argument, because they do not change the
products to be supplied by researchers whose underlying quality of knowledge as public, they
integrity in conducting each separate research only change the way our legal systems sanction
project is not under question. The problem raised its use.)
here does not result from any individual Second, knowledge is public in a proprietary
wrongdoing but rather centers on a robustly sense. That is, the public owns it by virtue of
collective effect of individually blameless acts (1, having paid for its production through taxes. Not
2). only do public grant funds pay for much research
directly, there are also many implicit forms of
What is the Problem? subsidy that enable scientific education and
Those who accept the analysis so far will practice—the public school and university
concede that the market for research funding system being only one large example.
affects the distribution of research products; Third, knowledge is public in a normative
however they may still deny that this is an ethical sense. We pursue research as a calling—as
problem. For instance, if one subscribes to the something we do for our fellow humans—as
“marketplace of ideas” model of truth (cf. 3, 4), much as for our own livelihood and reputation.
then a free market for ideas, for their sponsorship The cobbler usually does not take up this trade so
and dissemination—such as has been that the feet of the world may be shod, but
described—is the most efficient system for researchers often are motivated by a desire to
allowing the truest views to emerge. As long as contribute to the progress of humankind’s
each seller and buyer of ideas is free to make her knowledge. Most of us believe that knowledge
own choices for her own reasons, the invisible exists to serve society or humanity, not only for
hand of the market will guarantee that the best the “consumers” who pay for the production of
(i.e., the most sought after) ideas flourish. If an research.
area of research truly has merit, surely some A free market of interactions between
clever grantor will see that there are returns to be purchasers and suppliers of research (or any)
obtained and enter the market. This model products might perhaps optimize the satisfaction
presupposes that within a free market for of direct parties to these transactions. However,
research funding, the best quality science will the interests of the public are not directly
receive the best funding simply by virtue of its represented in reaching this theoretical market
quality. equilibrium. A bias away from the public interest
Two different rebuttals to a marketplace of will result, to exactly the extent that research-
ideas model are offered here; each based on an demanding grantors and the broader research-
accepted standard for assessing the inherent using public have systematically different
quality of research products, independent of interests.
market demand. The first argument is
democratic, while the second is elitist. The elitist argument: good science is an
autonomous pursuit
The democratic argument: knowledge is a A body of scientific knowledge is not simply a
public good collection of individual researchers’ products. It
Although effective demand for research is is produced by a community of scientists.
exerted by grantors, research products do not Individual researchers may have unconscious
serve only grantors. Knowledge is a public good biases (5) and may certainly commit honest
in at least three different senses. errors. These flaws can only be corrected from
First, knowledge is public in the technical another’s perspective. Thus the quality of
economist’s sense: knowledge products are often scientific knowledge emerges from interaction
non-excludable or offer positive externalities to among knowledge producers, not only from the
people other than the purchaser. Research quality of any one producer’s activity. This self-
products are not only there to be used by a correcting feature of scientific knowledge is
317
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
historically traced to the work of Herschel, Whitehall studies in Great Britain showed that
Merton and Popper (6, 7, 8), but the motif of a age-adjusted mortality from nearly all causes
self-correcting, autonomous body of science- varied inversely and quite significantly with civil
producing experts is also implicit in Kuhn’s service grade even when controlled for
classic account of progress through revolution individual health variables such as smoking. In
and in post-Kuhnians such as Laudan (9, 10). If other words, the higher the civil service grade,
one subscribes to any such elitist model, the the less likely these civil servants were to get ill
proper advance of scientific knowledge results or to die, all other things being equal. Similar
from the intellectual judgements made by a relationships between social status and
community of qualified researchers, not from the biochemical health indicators have been found in
economic demand for research. If aggregate experimental monkeys (12).
demand for research does not correspond to the The nearly exclusive emphasis on one or two
range of projects that researchers would choose modes of causation is problematic because the
to pursue on solely intellectual grounds, then to others might equally and perhaps more cheaply
this extent, the body of knowledge being lead to better public health. If prevention is
produced will exhibit a form of bias. intrinsically better than cure, then controlling
large scale correlates of disease is better than
Why does collective bias matter? using genetic or pharmaceutical technology to
Ultimately, the main reason we care about treat disease. To make this concrete: a breast
integrity of research at the individual level is that cancer gene may be significantly correlated with
the intellectual adequacy of a body of research is breast cancer, but possibly not more so than
vitiated by research corruption. Corrupt practices poverty, radiation, or other environmental and
produce dubious, misleading results. From either economic factors. If the public and policymakers
a democratic or an elitist perspective, we should become aware of the first relationship but few
care about collective bias for exactly the same researchers are pursuing the rest, a misplaced
reason—because a body of research formed by emphasis will be put on genetic therapy and too
demand may mislead researchers, students, the little effort on other possible methods for
public at large, and policymakers. In any field addressing this disease.
based on multi-causal or probabilistic systems, As long as there is a predominant demand for
the problem of collective bias resulting from the genetic research, we will continue to get genetic
demand for research should be of particular results. What is more, a disproportionate number
concern. of apprentice researchers will continue to be
trained in the area of genetic medical research
Case: Causes of Disease (not environmental or social medicine) and to
Sylvia Tesh remarked in 1988 that studies based develop a commitment to being geneticists rather
on a contagion model of disease were best than some other kind of health researcher. They
funded, most prestigious and generally dominant in turn will have incentives to conduct and to
in American medical research (11). Today (in support future medical research on a genetic
2001) contagion has been joined or perhaps model. Thus demand is not only affecting
displaced by genetics as the dominant cause of research in the present, it is also influencing the
disease to be researched. A third model shape of the future research producing
underlying research studies is lifestyle theory, the community.
idea that modifiable personal behaviors result in
illness. All three of these causal models fall Why is it an Integrity Problem?
under an overarching individualistic framework, If the analysis of the paper is accepted, then the
where disease is located within the person, demand for research poses some kind of social
whether in her genes, in a viral or bacterial agent problem. Yet as an ethical problem it is
she has taken in, or in her choice of (un)healthy paradoxical because we cannot find the
behaviours. By contrast, environmental, wrongdoer. For this form of research corruption
economic and psycho-social causes of disease to arise, there need not be any demand for
receive far less attention (and far less funding). corrupt research nor any suppliers of research
Evidence from other First World countries who are willing to be corrupted. No personal
suggests that these would be highly fruitful areas misconduct or violation of individual research
of inquiry. To take only one instance, the autonomy needs to take place. There must only
318
––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Gowri, A Market Approach to Research Integrity
be a situation where funding organizations freely historical and international comparisons,
select the type of research they will fund from qualitative social studies of market effects on
among various projects and models being mentoring and career choice, or quasi-
proposed. In other words, corruption of research experimental studies of factors involved in
due to the demand for research is a robustly research problem choice, for example. Finally, I
collective problem; it is not a problem that can be do not expect it will be easy to fund research
resolved by making individual people behave about collective market effects on research
more honestly or fairly. The reader may wonder, integrity, since funding agencies can hardly be
therefore, whether this is actually a problem of expected to have an interest in demanding this
research integrity, or just some kind of market kind of knowledge that would, after all, challenge
imperfection or political problem. The reply to their own role in directing the course of
this last objection lies in the professional status knowledge production. Such research would,
of researchers. however, offer valuable insight to the research
professions and to the public.
Research is a profession
Professionals are characterized by most ethicists Bibliography
as the bearers of many social privileges including 1. Niebuhr R. Moral man and immoral society. New York:
a monopoly on legitimate practice within their Scribner’s Sons; 1932. Reprinted; 1960.
domain, control of entry into that domain, and 2. Gowri A. Towards a moral ecology: the relationship
evaluation of one another’s competence (13). between collective and human agents. Social
Epistemology 1997;1(1):73-95.
Following this definition, scientific researchers 3. Wright JS. On courts and democracy. Miller A, ed.
are professionals. In exchange for their Greenwood Press; 1984.
privileges, the members of a profession are 4. Gowri A. Speech and spending: corporate political
collectively responsible for the character of their speech rights under the first amendment. J Bus Eth
practice as a whole: they must ensure that it 1998;16:1835-1860.
benefits a society as much as possible, and at 5. Barnes B. Interests and the growth of knowledge.
least that it does no harm. If researchers are Routledge & Kegan Paul; 1977.
professionals then they are not only responsible 6. Herschel J. A preliminary discourse on the study of
for doing research honestly, they are also natural philosophy. 1830. Reprinted: University of
Chicago Press; 1987.
custodians of their realm of research. Collective 7. Merton R. The normative structure of science. 1942.
responsibility of this kind has been accepted by Reprinted in: Merton R. The sociology of science.
traditional professions including medicine and Storer A, ed. University of Chicago Press; 1973.
law, and by many newer ones such as nursing, 8. Popper K. The logic of scientific discovery. Harper &
accounting and insurance (14). Of course Row; 1959.
researchers in a field may not be the only persons 9. Kuhn T. The structure of scientific revolutions.
responsible for the collective integrity of that University of Chicago Press; 1962.
field. 10. Laudan L. Progress and its problems. University of
California Press; 1977.
11. Tesh, S. Hidden arguments: political ideology and
What can be done? disease prevention policy. Rutgers University
In this paper I have called attention to a type of Press;1988.
failure of research integrity that has not yet been 12. Wilkinson R. Health inequalities: relative or absolute
addressed in research on research integrity. I do material standards? Brit Med J 1997;314: 591-5.
not pretend that it will be easy to address the 13. Airaksinen T. Professional ethics. Encyclopedia of
problem of collective integrity in knowledge Applied Ethics 1998;3: 671-682.
production: indeed, intrinsically collective 14. Horn R. On professions, professionals and professional
problems tend to be philosophically and ethics. American Institute of Property and Liability
Underwriters.
practically difficult (cf. 2). However, just
because a problem is not easy to fix, this does not
mean we should ignore it.
The existence argument for market effects on
the integrity of research must be supplemented
with research on the magnitude of these effects.
Such empirical studies could document the effect
of demand on research programs through
319
Methods for Research on Research Integrity: Doing Research on Sensitive
Topics
Margot L. Iverson, Department of History of Science and Technology, University of
Minnesota, Twin Cities, USA; formerly with the Program on Scientific Freedom,
Responsibility and Law, American Association for the Advancement of Science, USA
Mark S. Frankel, Program on Scientific Freedom, Responsibility and Law, American
Association for the Advancement of Science, USA

Keywords: Academic environments, Deviant behavior, Evaluation research, Methodology, Science

careers, Scientific misconduct

Promoting research integrity requires a greater understanding than we now have of the factors that
influence the full range of research conduct. There is a dearth of empirical research addressing issues
related to research integrity and misconduct in science. It is critical, therefore, that more research on
these issues be supported, not only to provide useful guidance to researchers and to the formulation of
appropriately measured policy, but also to stimulate a critical mass of scholars to develop research on
research integrity as a legitimate field of scientific inquiry. Such research must employ rigorous
research designs and methods of evaluation.
The “Session on Methods for Research on Research Integrity,” co-organized by Mark S. Frankel
and Felice Levine, considered the methodological challenges faced by researchers studying research
integrity and discussed research approaches best-suited to this topic. Four speakers presented
different models and strategies for conducting research on research integrity and suggested promising
areas for future research. The session concluded with discussion of a possible research agenda for
research on research integrity. This account is a summary of the session.

Contextual Effects in the Study of Academic Misconduct

Melissa Anderson, Associate Professor of Higher Education at the University of Minnesota, presented
conceptual models of scientific misconduct that could be used to guide research on the role of the
academic environment on research misconduct. Studying different aspects of the research context in
which incidents occur can move researchers away from focusing on prevalence, which is difficult to
determine and of limited utility, to examining other research questions useful to institutions trying to
promote research integrity. Researchers face several methodological challenges in investigating
research misconduct. Misconduct is a sensitive topic that individuals wish to keep hidden from
researchers (and others), making it hard to observe, and incidents are relatively rare, making them
difficult to find and compare. The academic context in which misconduct is to be studied also can
create methodological difficulties. Research areas in which perpetrators of scientific misconduct

Corresponding author: Mark S. Frankel, Program on Scientific Freedom, Responsibility and Law, American Association for
the Advancement of Science, 1200 New York Avenue, NW, Washington, DC 20005, 202-326-6793 (voice), 202-289-4950
(fax), [email protected].
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
function can be very technical, requiring time presents many possible research projects.
investigators to possess some mastery of the The contextual influences of the broader research
specialized subject matter (or to collaborate with environment on these four stages, from such
someone who does). Another problem can be sources as disciplinary societies, journals,
the autonomous nature of academic researchers, industry, government policies, and elsewhere,
which makes their behavior difficult to observe also suggest many useful research topics.
or to confirm independently. Additionally,
research integrity research is not always welcome Scientific Misconduct as a Form of
by institutions or departments, out of fear of Deviant Behavior
media or legal attention, and individuals and Researchers who engage in scientific misconduct
organizations may not cooperate with are behaving in a presumably deviant way that
researchers. violates both legal and social norms. Conducting
Rather than artificially disassociating empirical research on research integrity and
misconduct from research, conceptualizing it as misconduct therefore requires that researchers
linked to unavoidable research error is one way consider the implications of studying deviant
in which misconduct can be understood in the behavior in designing and conducting their
context of the research process. Error and research. In her presentation, Eleanor Singer,
misconduct both involve issues of intention and from the Institute for Social Research at the
acceptability, with misconduct being both University of Michigan, discussed some
intentional and unacceptable, and inadvertent methodological considerations arising from this
error being the reverse—acceptable and understanding of research misconduct as a form
unintentional. Anderson identified two other of deviant behavior. In addition, she also
categories as well, avoidable error, which is presented some applications of more universal
unintentional but also unacceptable, and “minor research principles to research on research
hypocrisies,” which are intentional but integrity.
acceptable. Studying these categories of Deviant behavior is difficult to study because
avoidable error and minor hypocrisies, which there are strong incentives for both perpetrators
presumably are much more common than and the institutions at which it takes place to
misconduct, may provide information on the keep it hidden. This makes it difficult to observe
contextual influences on misconduct that is directly, and so researchers must resort to asking
difficult to obtain by other means. And since subjects to report incidents. Two of the most
intent is hard to determine, some instances of common methods used are self-administered
avoidable error may be incidents of misconduct surveys and interviews. These are more likely to
that have never been so identified. Other topics produce honest answers if the confidentiality of
for further research that grow out of this linkage those participating can be guaranteed. Surveys
between misconduct and error are how scientists that are self-administered, further ensuring
decide what separates misconduct from these privacy, also can improve rates of subjects’
categories, and if and how they deal with error as veracity. Another useful research method for
well as misconduct. some research questions is to present subjects
Another way to examine context is to with vignettes of ethical quandaries in research
consider not just the actual incident of and to ask them how the researcher in the
misconduct, but rather to understand cases as vignette would behave. Such vignettes are most
having four distinct stages: the context useful when the type of research and the status of
(institutional, disciplinary, and immediate lab) in the researcher in the vignette parallels those of
which the incident occurs, the misconduct event the subject, as this increases the chance that the
itself, the exposure of the misconduct, and the answer will reflect their own behavior.
consequences for the perpetrator and others. Vignettes also can be used to study what
This framework provides a way of considering behaviors actually are regarded as violations of
and comparing different aspects of misconduct so standards of conduct by members of a particular
that interactions between each stage can be field.
explored. For example, what impact does the Like other forms of deviant behavior,
context of funding sources and mechanisms have opportunities to engage in scientific misconduct
on incidents of misconduct? Longitudinal as well as opportunities for observing it can vary
research of patterns of interactive effects over depending on factors such as the discipline of
322
––––––––––––––––––––––––––––––––––––––––––––––––––– Iverson & Frankel, Methods for Research on Research Integrity
research and the size of a department. Also, Rosenfeld, a Professor of Sociology at the
motivations for deviant behavior may vary, based University of North Carolina, presented a
on incentives and reward systems present. sociological framework to consider how
Singer also presented several other principles scientists learn about ethical research practices at
of empirical research that are critical to different career stages, what it is they learn, from
producing rigorous empirical research on whom, and why sometimes they learn the wrong
research integrity. Researchers must establish the lesson (i.e., unethical behavior). At each stage of
questions they wish to answer with their a scientist’s career, several nested contexts
research. To obtain consistent answers and influence research integrity. In the immediate
meaningful results, terms used also must be research environment, researchers are exposed to
defined. For example, since norms and peers, mentors, teachers, collaborators, and
definitions of research integrity and misconduct students. Surrounding and overlapping this
vary, these terms must be clarified so that all immediate environment are the context of
researchers and subjects are using a standard department and institution and the broader
definition. If not, ambiguity may be introduced context of journals, professional societies, and
into the data. (Research that explores differences federal policies.
in norms and definitions of misconduct could be Rosenfeld discussed some potential research
very useful in helping to interpret current data on projects at each stage of a scientist’s career, from
prevalence.) The populations to be studied also undergraduate through senior scientist. Currently
must be selected so that comparisons can be available research on undergraduates has focused
made. When choosing research methods, the on the conduct of science students in the
match between method and research question classroom and has indicated distressingly high
should be carefully considered. Direct rates of plagiarism and fudging data. Are
observation, deliberate experimentation, advanced students engaged in independent
questioning subjects, and analysis of official research projects more or less likely to fudge or
records are all possible methods, and each has plagiarize data in the research environment?
advantages and disadvantages. The choice of This would be an especially interesting research
method also involves a selection of the indicators topic since those undergraduates who do
the study will use. Official records of complaints participate in research are more likely to continue
of research misconduct, for example, will yield on to graduate school than other students. For
different information about incidence than data graduate students, research has suggested that the
collected through surveys of bystanders or interaction between them and their mentors is
perpetrators. Since descriptive statistics are critical to their subsequent ethical behavior.
much more meaningful in a comparative context, More research is needed on how aspects of this
it is important to consider how different parts of a interaction affect the information on research
study can be made sufficiently equivalent so that integrity transmitted. The role of other graduate
data can be analyzed comparatively. The students, and the effect of isolation from peers on
research conducted by Judith Swazey, Melissa ethical behavior are other potential topics. To
Anderson, and Karen Seashore Louis on integrity what extent are graduate students who interact
issues in graduate education is a good example of frequently with their peers learning ethical (or
the effective application of these research unethical) behaviors from them? As researchers
principles to research on research integrity (1-2). move from being graduate students to post-
doctoral trainees to junior scientists, the broader
Influences on Research Integrity at research community context becomes more
Different Stages of Academic Science important. Journals and scientific societies may
Careers become more influential in shaping junior
Another research model that can be applied to scientists’ behavior. Do varying standards of
research on research integrity is the effect of the evidence adopted by different journals influence
academic environment on researchers at different researchers’ research practices? For example, if
stages of their careers. Although many scientists a journal requires that all underlying data be
take a class on research ethics early in their accessible, does that have an effect on the
training, the major influence on how they learn to accuracy of the researcher’s analysis of the data?
conduct ethical research is usually the How does the pressure to publish affect what and
environment in which they work. Rachel how researchers conduct research? Regarding
323
Proceedings: Investigating Research Integrity (2001) –––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
scientific societies, how does leadership on following programmatic and research design
research integrity from societies impact the elements need to be in place.
behavior of members? Do society ethics codes First, program goals must be clearly defined
and ethics prizes influence members? And for for a specified target audience (e.g., graduate
senior scientists, who are likely to become part of students will be made aware of the ethical
the leadership of societies and departments, how standards for research and the strategies for
do these roles influence their own research adhering to these standards). Second, activities to
conduct? achieve these goals must be designed and
Contextual questions exist for each stage of a implemented (e.g., an educational program
scientist’s career, and studying these questions consisting of a one-credit course is established as
can identify the conditions under which a graduation requirement; it is taught every fall
interactions in a particular context lead to the semester). Next, a plan for the evaluation of the
learning of ethical or less ethical research program’s implementation process and outcomes
practices. That researchers might receive mixed needs to be delineated, including measurements
messages from the different contextual and instrumentation (e.g., measures of
environments was noted by an audience member, knowledge using a paper/pencil test or measures
and Rosenfeld concurred, noting that some of decision making using case scenarios), timing
contextual messages may promote unethical of data collection (at the end of each course),
behavior and that it is important to assess how methods of analysis (quantitative), and format for
competing messages are dealt with by scientists. reporting the results and implications for an
Another factor to consider in research is how organization’s activities, since it is essential to
these nested contexts affect individual incorporate a system for linking knowledge
researchers in different ways. A researcher’s gained through research to organizational
gender, race, country of origin, or sexual planning and action.
orientation can all impact the individual’s Evaluation research assesses the overall
interactions with the surrounding environment. effectiveness of an organizational program and is
used to improve programming so that goals are
Utilizing Evaluation Research to Assess met and resources are used efficiently. It is based
Research Integrity Programs on an open system’s model of organizations
Joyce Iutcovich, President of Keystone (“open” because the organization is open to
University Research Corporation in Erie, political, social, and economic influences from
Pennsylvania, presented an overview of the the external environment). As conceptualized
contributions that evaluation research can make using this model, evaluation research provides
to research on research integrity. Along with evidence, which becomes part of the continuous
basic research, which addresses questions about feedback loop that constantly works to improve
causality and contributes to theory development, programmatic efforts. Ideally, programmatic
evaluation research provides the link between efforts that address issues and concerns related to
theory and practice. When research institutions research integrity are based on theoretical models
and scientific societies develop research integrity that provide an understanding of research
programs based in part on theory, evaluation integrity and how to ensure it within a population
research plays an important role in assessing the of researchers and scientists. Once implemented,
effectiveness of these programs. Further, it offers evaluation data on these programs are collected,
a system for transferring knowledge gained analyzed, and used for program improvement.
through research to program improvement efforts Evaluation research also provides another critical
over time. assessment of the theoretical model, which
Evaluation research is conducted within the establishes the framework for the program. This
context of social action programming. It focuses further enhances theoretical development by
on an assessment of the implementation process providing evidence about what works and what
as well as the outcomes for targeted groups. doesn’t work as predicted by a theoretical model.
Process evaluation determines whether a program
has been implemented as planned; outcome Session Conclusion—Developing a
evaluation determines the short- and long-term Research Agenda
impact of a program on the target group(s). To Felice Levine, Executive Officer of the American
conduct a process and outcome evaluation, the Sociological Association, addressed the scope of
324
––––––––––––––––––––––––––––––––––––––––––––––––––– Iverson & Frankel, Methods for Research on Research Integrity
research integrity and misconduct concerns, the The session ended with some questions and
challenges for undertaking study of such issues, comments from the audience. Among the final
and the need to attract researchers with broad comments was the observation that many of the
expertise. Also, synthesizing many of the topics presentations focused more on context than on
raised in the presentations, she concluded by individual behavior and that this seemed to
suggesting steps needed to establish a research reflect a shift from individual character to
agenda for studying research integrity. research context in understanding research
Prior to designing an agenda, the scope of the misconduct. The need to include “organizational
research and related topics on research integrity misconduct” in this field of research also was
and misconduct must first be determined. Along voiced.
with fabrication, falsification, and plagiarism, Studying research misconduct presents
issues of conflicts of interest, human research several kinds of methodological challenges,
participants, confidentiality, authorship including difficulties in observing deviant
determination, data access/sharing, data design, behavior and in conducting research in an
and accurate representations and interpretations academic environment. Researchers,
of data all may fall within this subject area. The institutional review boards, and funders must be
complexities involved in conducting research on sensitive to these matters and give due diligence
research integrity also must be considered. Since to research design and methods. Nothing could
deviant behavior is often hidden from outside set the field back more, even before it takes
view or occurs among powerful elites, there are shape, than sloppy, inappropriate, or poorly
many challenges to obtaining empirical data on designed or applied research methods.
research integrity. Political concerns within and
between organizations also may inhibit research. Bibliography
Also, since this research could benefit from 1. Swazey JM, Anderson MS, Louis KS. Ethical problems
research methodologies and frameworks from a in academic research. Academic Scientist 1993; 81
variety of disciplines, attracting researchers from (November/December): 542-555.
a broad range of disciplines is crucial. Across 2. Anderson MS. Uncovering the covert: research on
academic misconduct. In: Braxton JM, editor.
disciplines, important areas of expertise for such
Perspectives on Scholarly Misconduct in the Sciences.
research include history and sociology of Columbus, Ohio: Ohio State University Press; 1999. p.
science; work, occupations, and professions; 283-314.
research ethics; deviance and white collar crime;
decisionmaking; and organizational behavior.
Levine then presented initial steps to be
taken to establish an agenda. The stakeholders in
research integrity—including the individual
investigators, research teams, scientific societies,
potential funders, subjects to be studied,
policymakers, and the public—must be
identified. Data sources already available from
federal agencies and other organizations as well
as resources needed but not available should be
assessed. Funding sources and mechanisms
should be identified, and structures—including
conferences, working groups, panels, and large-
scale collaborations—should be set in place to
provide frequent opportunities for scholars to
communicate. Finally, to develop a community
of researchers working in this area, a substantial
investment is needed to provide educational
opportunities for researchers from different
disciplines and at different career stages. These
opportunities could include internships for
students, postdoctoral and mid-career incentives
or awards, and specialized training programs.
325
Research Misconduct: A Multiperspectival Approach
Robert J. Silverman, Fielding Graduate Institute, Santa Barbara, CA, USA

Keywords: Scholarship norms, Social science misconduct

This paper engages the topic of research integrity with a foundation that differs from that which
appears in more traditional treatments of misconduct (1, 2). These other treatments are grounded in
the notion of “role;” that is, scholars are in role and guided or controlled by certain norms whose
abrogation are role sins.
But what if one, at base, defines the scholarly life in different terms, not on the basis of distinct
research practices, mandated by professional role, but where there is no separation between how one
lives one’s life and how one produces as a scholar? Or, what if one uses one’s scholarly work as a
way of living one’s life in the broader society and culture such that one’s impact (3) will reshape the
environments in more acceptable terms to oneself and others? Current misconduct considerations
need to pre-suppose social and psychological patterns under girding professional life that are, in fact,
more varied than often assumed.
The traditional norms, such as communality, universality, organized skepticism, or even their
opposite as counter norms–which continue the salience of the dominant terms–and which have
historical grounding (4) do not capture these definitional differences, as the wars among the
epistemologies make rather clear. How might we engage this thicket? We need to do so less as
watchdogs, which is the current preferred pattern. Charging and defending individuals and social
institutions is but one set of approaches. Holding individuals to norms that are irrelevant to their
definitions of scholarly work and life will not motivate us to consider how we might address the heart
of the knowledge creating enterprise in its foundational diversity. This is a time for rethinking what
norms and standards should obtain, and we need to do so with energy and with expectations of their
significant complexity.
Unfortunately, this complexity and inventiveness is not apparent in the Report of the Commission
on Research Integrity, entitled Integrity and Misconduct in Science (5), though it invites definitions
from various single and multidisciplinary fields. It argues that:
Research misconduct is significant misbehavior that improperly appropriates the intellectual property or
contributions of others, that intentionally impedes the progress of research, or that risks corrupting the
scientific record or compromising the integrity of scientific practice....(5)
It develops new terms, in place of “fabrication, falsification and plagiarism,” namely,
“misappropriation, interference, and misrepresentation” (5). The first includes plagiarism and also
notes the improper use of what is essentially confidential or privileged information; the second covers
the damaging of others’ research property; and the last deals with attempts to deceive, from omission
or commission. There are other forms of “professional misconduct” added to that of research
misconduct, primarily dealing with attempts to obstruct investigations of misconduct. In addition,
there are calls for both academic institutions and professional fields or disciplines to develop codes of
Corresponding author: Robert Silverman, Fielding Graduate Institute, 2122 Santa Barbara St., Santa Barbara, CA, 805-687-
1099 or 360-566-8188 (voice), 360-566-1101 (fax), [email protected].
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
conduct and to provide a variety of educational Many scholars evoke images suggested by
experiences for scholars regarding ways of the distinction between newly conceived and
behaving professionally to avoid research established knowledge, including Lorraine Code,
misconduct. While this committee’s initiative in who in What Can She Know? (6), extends
rethinking misconduct and its invitation for constitutive and regulative patterns to the
contingent organizations to focus on “local” community, as well as to the knowledge that
concerns appears positive, there appears to be no transpires among its members. It is the
larger vision in the recommendations, apart from interaction of various approaches to knowledge
the Committee’s observation that a basic and to community that shapes the academic
principle was “that scientists be truthful and fair” contexts in which scholars live their lives. Let us
(5). That is, there is no sense that inquiry, the use Code (6), though one could focus on many,
actions of scholars, or the codes of disciplinary/ many others to make the same distinctions under
multidisciplinary communities vary considerably. the banners of their foundational goals (e.g.,
However, these are not just local factors that are 7,8,9,10).
at issue, but different world views, different
logics-in-use. Unless our codes take them into Academic Communities
account, scholars will be applying the wrong First, let us differentiate among “academic
rules and norms for their own and others’ communities.”
research and ignoring the need to create norms 1. In a constitutive community: “...every
that have relevance. cognitive act takes place at a point of inter-
section of innumerable relations, events,
Communities of Scholarship
circumstances, and histories that make the
The focus of this paper is to begin generating
different sets of values or norms that have knower and known what they are, at that
relevance for alternative ways of knowing, not time... (It focuses on) the complex network
reflected by singular methods or in specific of relations within which an organism
fields, but in light of co-existing epistemological realizes, or fails to realize its potential... (6).”
contexts. And, the location is the social sciences The community allows for interrogation,
and not the natural sciences—which are assumed renegotiation ; it evidences trust which
in nearly all conversations regarding research “involves making oneself vulnerable... (6).”
integrity. We do need to refresh the conversation 2. In a regulative community: One sees
with creative possibilities that are in keeping authoritarian knowers who “...claim credibil-
with how and with whom scholars do their work.
ity on the basis of privilege alone or of
As these alternatives become more evident there
is a responsibility to consider their implications ideological orthodoxy... (6).” Code suggests
for normative as well as other issues. Alternative that there is an obsession with autonomy and
warranties for different ways of doing an overemphasis of the self (6).
scholarship and being academics are needed. Regulative communities have historical
Three years or 36,000 miles is just one solution. dimensions with regard to the participation of
various actors. These dimensions have
Academic Contexts significance for an understanding of current
This paper differentiates the notion of “academic professional directions, as exemplified by Hull’s
context” in terms of the generativity or work on the consequences of members’
embeddedness of the knowledge and the contributions to the field of taxonomic biology.
openness or stability of the communities within Regulative fields are more integrated and
which scholars work. It appears that the concepts planned with professional divisions than are
of “constitutive” and “regulative” can be constitutive communities which are more
appropriately applied as significant adjectives to organic, emergent, and fragmented, with
both knowledge and community such that we collections of individuals coming together for
identify working cohorts of scholars who are cause, that is, the attraction of a problem or issue
engaging with knowledge differently and whose rather than a continuing research or theoretical
work will have differential value to other focus.
stakeholders who identify with each other. Much of academic life is geared to the
celebration of success, and the study of action, in
328
––––––––––––––––––––––––––––––––––––––––––––––––––– Silverman, Research Misconduct: A Multiperspectival Approach
a regulative model of community, is reflected by example, on a common-sense basis, it is clear
research prizes, citations studies of a that the field of psychology embraces humanistic
contribution’s value or impact, and the life of and behavioral alternatives, as well as areas that
research schools (12). Constitutive communities are laboratory and clinically-based. To place a
often are inappropriately placed in the regulative field such as this conceptually in one area is to
frame as when one discusses a field or area as deny its multiplicity. It contains cohorts of
being “pre-paradigmatic,” as if it were “pre- scholars who could be differentially placed in the
pregnant.” It occurs inappropriately when a four-fold scheme, as suggested above. But, such
comparison of citations patterns is made between locating is not necessarily a “cold-blooded” act
physicists and educationists, as if research papers of placement; it occurs as well among scholars
are not used by the latter for policy guidance, a who react with some emotion to each other’s
value that goes unrecorded, and assume the only work. This paper acknowledges the legitimacy of
value of a paper is in crafting new research. Now all that claim to be knowledge communities and
let us differentiate among kinds of knowledge: asks the reader to surrender her or his current
1. For constitutive knowledge: one takes categories for the suggested set of alternatives.
account of testimony and cognitive interde- Unless one is so willing, it is possible that
pendence (6),...” letting ‘objects’ of study researchers may be locked into more narrowly
speak for themselves..., ...understand(ing) defined debates than relevant. In commonplace
language, this is seeing the trees and not the
difference and accord(ing) it respect (6). It
forest.
grows by accretion without a preexisting As suggested, then, there are four knowledge
frame. contexts which are the relation of constitutive
2. For regulative knowledge: there are more and regulative possibilities for both the
standard forms; it is more hierarchical, is community and the knowledge its’ members
informed by such principles as objectivity produce.
and value-neutrality at the same time it is 1. Scholars in regulative communities can
also more adversarial and territorial (p. 6). develop regulative knowledge. This is the
Constitutive knowledge is developed from many traditional context in which fields develop
sources; it is constructed piecemeal, whether the incrementally within well-established paradigms
source materials are concepts, ideas, or data or theoretically/empirically informed schools.
bases, or some combination derived from While these areas grow in terms of the
resources made proximate by the scholar who “agencies” of various human and machinic
creates such a bundle to address a “problematic.” components, the goal is the stabilization of
A marvelous example is reflected in Shapin’s knowledge (14) by persons who, with a
work (13) where he writes in his introduction: complementary set of scholarly interests, seek
“(This book on social history) is concerned with answers to research questions of acknowledged
questions ...which have traditionally been the importance. This is the context that Kuhn and
preserve of philosophers; it uses evidence and most other commentators assume, and which
techniques customarily owned by historians; and they unfortunately assume to be universal.
the conclusions it arrives at are broadly 2. Scholars in constitutive communities can
sociological in form and substance.” He also develop regulative knowledge. This occurs
disavows an interdisciplinary orientation and when cohorts of scholars in a variety of fields or
hopes to be identified as an historian. Regulative areas center around the work of theorists or
knowledge is more standardized, cumulative, schools of thought that pull them into similar
specialized, and stable. logics. For example, persons in a variety of
There is an interaction between “knowledge” “disciplines” or areas study Piaget, or Kohlberg,
and “community,” that is, alternative kinds of or Kuhn, or critical perspectives as developed by
knowledge are crafted in communities of either Foucault. Concepts or theorists are treated in
type. It is important to not necessarily equate the canonical fashion as they are “applied” to various
resulting four types of possibilities with “problematics” by persons in various locations.
disciplines or fields of study. Some fields or 3. Scholars in regulative communities can
disciplines may have contexts of only one type; develop constitutive knowledge. This occurs
there are many in the social sciences that when researchers in an area, such as higher
embrace those reflecting multiple logics. For education or the sociology of science, study an
329
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
issue by bringing together unique knowledge both current and historic examples of misconduct
resources, that is, concepts, theories, and by figures who made foundational contributions
methodologies from a number of sources to (e.g., Newton, Pasteur), contributed significant
frame and study the concern at hand. The focus is work (e.g., Burt), and did normal science, all of
on the question, with the inventiveness of the which are treated in historical studies and
scholar addressing others who examine similar journal, news and list-serv accounts; admonitions
questions from very different resources and and suggestions regarding how the mechanisms
choices. of the scientific community (e.g., editors and
4. Scholars in constitutive communities can peer reviewers) can act to be aware of
develop constitutive knowledge. This occurs misconduct reflected in submissions; how
when individuals from a variety of locations various stakeholders (e.g., lawyers, scientists)
come together to establish evolving focus on different dimensions of misconduct and
understandings in a particular way. As an the actions of institutions (e.g., universities) in
example, one might have feminist scholars such instances; the debates regarding what
develop an understanding that has both misconduct includes in practice and the
independent and interdependent sections on ways delimiting of the practices that are so situated, to
of doing professional work, such as teaching. include the on-going conversations sponsored by
The community of interest and understanding are such organizations as the American Association
organic, evolving, and this is reflected in various for the Advancement of Science that provide for
patterns of intersection. public reflection on the issues. There are many
In earlier work, this researcher has attempted voices, from the philosophers, historians, and
to appreciate the meaning of these differences in sociologists, to the aggrieved parties and those
terms of their implications for electronic who are situated in different parts in the various
publishing and to examine how they shape the dramas; to those who police the science
nature of argumentation within and between community.
scholars who live in these academic contexts (15, As one can note, the concern is primarily for
16). There is work that establishes these the misguidance of the scientist, and there is
distinctions through philosophic attention (e.g., limited attention to the community of science as
17, 9, 10). It can be argued, from a broad playing a part in the perpetuation of misconduct.
multidisciplinary base, that these distinctions are While the Public Health Service Report reflects a
primary ones and that the particular labeling of concern for the role of universities and
the contingencies in this paper reflects professional societies and focuses attention on
alternatives that have universal meaning, though the “whistleblower” as she or he is treated by
other scholars have used different language and professional peers, the attention is on the
examples in their areas to denote these producer of knowledge as an individual and not
alternatives. The question for what follows is, elements of the community (5).
“What are the implications of alternative For example, if a journal editor sends a
production in different kinds of academic manuscript out for external review to peers who
communities for the meaning of misconduct?” If do not share the methodological bias of the
the community is attempting to appreciate what author or the reviewer comments on the paper in
comprises misconduct one needs to understand ways that suggests that he or she challenges the
the nature of the production within their logic of the approach to inquiry used by the
communities for such an understanding to have author, this would not be considered misconduct
value. by either the editor or reviewer. Rather, it would
be considered as poor judgment or bias, and the
Scientific Misconduct Pluralised actors would have nothing to answer to in any
Regulative Community/Regulative Knowledge. special forum. A journal that fails to publish its
This context accounts for the observations made policies and fails to send manuscript reviews to
by those who are part of the historical and authors would not be cited for “killing” the
contemporary conversations with which we are scholar by denying opportunities for access.
familiar. Trust, not justice, is the focus of the operant
It is a voluminous set of materials which norms in this fully regulative sector, and the
focuses on such domains as the traditional norms regulative community has little to answer to
of science and adherence to them in practice; under this additional value. Interestingly, there is
330
––––––––––––––––––––––––––––––––––––––––––––––––––– Silverman, Research Misconduct: A Multiperspectival Approach
a vital literature on how academics “cool out” or notions, or meanings, one is repeating
their colleagues, either because they produce on material, replicating others’ language in which
the “margins” of accepted knowledge or because the constancy of the foundational work is
of gender (18). Practices of disciplinary bodies repeated. If one uses the term “needs” or
or forums, and scholars associated with them, are “paradigm”, then the source is known, not only
not considered as having relevance for in terms of the original authors, but the wall of
attributions of misconduct. The actions of a support behind the term. Those in the various
majority or those who operate from power are fields who are using canonical concepts or
hardly ever placed in such arenas. theories would find little value in using the work
Unfortunately, as well, the work on of scholars in other fields who use the same
misconduct that has relevance for this context is literature, say by Piaget, since, it would have no
assumed to be of value more generically. This applicability to their work in question. There is
allows us to ignore what is not normative in the a “catechism,” then, which like any canonical
other approaches to knowledge construction. text comes from well recognized sources and
Constitutive Community/Regulative which needs repeating, which migrates, as the
Knowledge. As noted above, in this context parameters it asserts are firmly supported by
scholars from various disciplines, forming a peers. Growth occurs as additional applications
colleague group, typically focus on the work of and connections among key spokespersons are
particular theorists, or the implications of articulated and as peers defend and repair
particular macro concepts or worldviews. For boundaries that support the difference “it” makes.
example, this would include scholars from such In the heuristic spirit of this paper,
fields as psychology, sociology, political science, misconduct can take a number of forms among
education, social work, and others, engaging with academic “true believers.” First is the failure of
the work of, say, Foucault and referencing the advocates to examine first principles. It is
colleagues who are similarly involved, rather suggested here that it is inappropriate for a
than colleagues who share a disciplinary-derived scholar to support and defend certain points of
designation, such as deviance in sociology, or view without giving serious consideration to the
analytical or mathematical geography. origins and consequences of the point of view or
It is of interest that the constituents are scheme and to appreciate it in relation to
committed to certain ideas that often are in alternatives. Presenting a persuasive or defensive
defensive contention with alternatives held by case in reference to one’s advocacy is a
another cohort of scholars. Peers engage in foundation that can be avoided by simply
discourse around which one finds a consensus, acknowledging the value of a work and
with argumentation around the fringes as identifying oneself through such association. It is
implications and new application of well done frequently, often by graduate students and
understood and frequently articulated ideas are then by those who find the pull of the network, or
evoked and borders are defended. What does it invisible college as less a call for excellence than
mean to plagiarize when proselytizing and self- a sinecure for the privilege of self-evident truth.
affirmation are the orders of the day? The truth Blind advocacy here is no less significant than
of work rests on a self-evident body of original being a true believer in any other context, and it
conceptions and supporting material, such as that has less a place in the academy than in other
surrounding the Jungian foundation of the social institutions. It should be expected that
Myers-Briggs personality instrument. Advancing one’s agency and voice be earned, not by
a set of concepts or worldview is the challenge, mimicking a rhetoric that one finds attractive, but
and the use of additional supporting material by being able to articulate with others, especially
only aids the cause. This is not to say that it is those who reflect alternative perspectives, a logic
appropriate to use someone else’s work as one that one has reasons to represent.
own. But, unlike the scholarship emerging from The second concern is about the nature of the
the previous location, this work is not so much argumentation that ensues. One may try to
the careful extension or articulation of a body of besmirch the other, to insult and shame the other,
work growing on the edge of a well-prepared rather than challenge her or him on appropriate
community, as it is the “answer” for viewing a grounds. My study examining patterns of
concern. That is, since individual scholars evoke advocacy and defense in this context applied the
in their language certain commonly-noted truths, labels “contention and fortification” to describe
331
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
the interactions across schools of thought around of a topic by bringing theories, findings,
the same issues (16). At times, when reading the concepts, and methodologies together that will
various protagonists, one could call forth an enable her or him to understand an issue in a way
analogy from the larger society, namely, the no- that is deemed appropriate for the stakeholders
holds-barred punching and kicking strong-person involved. The titles of two articles, the first in
contests that seem to have captured some Configurations: A Journal of Literature, Science,
interests among fans in the larger community. and Technology—(a journal title that itself makes
Since one’s identity is a function of “a point of the point)— “The Pathology of Painting:
view” or the viability of a particular body of Tuberculosis as Metaphor in the Art Theory of
work, one finds scholars needing to defend it by Kazimir Malevich” and in Social Studies of
any means available. Some of the strategies Science—”Literary Genres and the Construction
include attributing motives to the other that she of Knowledge in Biology: Semantic Shifts and
or he has not articulated; finding fault with the Scientific Change” are exemplars here. It is not
other because he or she did, or might have, read unusual for a person to claim that he or she is a
certain authors which the writer holds in multi or interdisciplinarian, or some variant
disrepute; using hostile words in combination thereof. Typically, the questions being
such that the rhetorical impact transcends their investigated are not of state-of-the-art
logical significance; and claiming others have significance to a subset of the academic
alternative perspectives even when they gave no community, but instead better reflect the unique
voice to such. The writer will allow the other interests of the involved scholars and present
view to survive, but in rather tender shape, their special and idiosyncratic ways of dealing
because it is easier to support oneself by arguing with approaches to questions that might be
against another view than doing so without an shared with peers.
enemy. Simply put, members of constitutive Certainly authors can engage in falsification,
communities working with and articulating fabrication, and plagiarism, and possibly with a
regulative knowledge should make judgments on lesser chance of being found out than one might
evidence and need not unfairly represent their were one in the fully regulative context. There is
positions or challenge alternatives or prevent less concentration of similarly educated and
their articulation. concerned peers and a greater variety of folks
It can be argued that the content of the who have different backgrounds and who roam
regulative knowledge at issue is of some the literatures and methodologies in seeking
importance. There are academic “belief” systems “fits” with their disciplinary and value-
that create inequities for others. Since theory constituted dispositions. What in a more specific
often outstrips its empirical base, or creates the sense does this suggest regarding misconduct?
possibility of work that can only support the It is suggested that not recognizing and
contentions of those involved, the impact of dealing with one’s own constructions as
theoretical systems on life worlds needs constructions is “misconduct.” That is, it is
attention—ideas which are removed from serious anormative to consider one’s own construction as
possibilities of destruction and upon which whole beyond reflexivity. Unlike work in the fully
careers are based. Is it a matter of misconduct if regulative environment, here there is no historical
one’s scholarship creates obstacles to others’ and progressive justification for a line of
human rights—either in terms of their status reasoning. Treating a solution to a problem as
possibilities or relative success? An example of self evident, such that it is not justifying itself in
such an obstacle would be the use of a relation to alternative treatments to the same or a
Caucasion-normed personality instrument and similar problem, does not allow others to
associated theory to classify the development of appreciate the added value that may accompany a
African American students, especially when an way of understanding. So, while one can gauge
instrument based on African American student the meaning of a work in the fully regulative
growth and change is available. We are context by its specific and its particular use of
responsible for the worlds we create. references, in this third context, such is not
Regulative Community/Constitutive possible. There needs to be the willingness to
Knowledge. In this context, scholars from justify the connections among the elements of a
regulative communities, such as from within the work and to engage in academic conversations,
sociology of science, develop an understanding which interestingly enough, are regular features
332
––––––––––––––––––––––––––––––––––––––––––––––––––– Silverman, Research Misconduct: A Multiperspectival Approach
of Social Studies of Science. Of course, others develop in conjunction. There is a “collective
need to be willing to engage, and this suggests integrity,” to use Mary Ann Caws’s term, with
the need for relevant regulative communities to the consequence being greater illumination,
have a meta-language available so that persons which leads both to interactants’ ensembling and
from different vantage points can engage each external stakeholder mutual appreciation. Since,
other. misconduct was originally defined within a fully
One might also suggest that it is cause for regulative context, one might wonder if it has any
concern, that is, misconduct, when the scholar applicability at all in this fully constitutive one
uses the work of others in ways that (19).
fundamentally change the elements such that the Individuals in this context reveal themselves
sources of origin would likely object to the through conversation, narrative, anecdotes,
implications of the uses of the work should the personally situated histories, and engaged
source have potential voice here, or when such professional settings. They do not use rhetorical
objections are ignored. This use could reflect a strategies to persuade, but attempt to present
number of possibilities, from the location of a themselves as evolving, with the risks that such
work, that is, where it is used, and how the confession might have. They reveal how they
sources are modified in a new treatment. For have grown or changed through various
example, today there is a major concern in the encounters with persons with whom they engage
management literature on the concept of in a scholarly way or with situations that provide
“resilience.” Some relevant questions might be, environments for learning and action. It would
“How can leaders have more of this?” “Do seem that it would be a misconduct situation if
successful CEO’s have this attribute?” In one used the information so revealed for private
approaching these types of questions, a scholar gain, either as the author or the audience. One is
uses narrative accounts of those who expressing or bringing forth an emotive/affective
demonstrated resilience for their survival: connection to the rational material; values are
returning POW’s and those liberated from Nazi clearly articulated, and a kind of privileged
concentration camps. It can be argued that one relationship is being established among the
dishonors the events and the lives of those who parties. While not referring to confidentiality in
perniciously suffered to use their accounts as terms of content, here one should maintain
conversation pieces for a cocktail party, and to confidentiality based on one’s respect for the
allow the CEO to think that he or she walks in person who has revealed something that is
the POW’s shoes. Such use reflects back on the “personal.” Thus, to ridicule the narrative of
original stories. They are re-storied, and another or to suggest one’s superiority in terms of
arguably, in ways that lessens their deeper intelligence, motives, or values is antithetical to
meaning and the meaning of those who found the orientation of this knowledge community. It
themselves as unwilling participants. freezes the logic of interaction and unnaturally
It is also a concern when scholars do not shapes the content of the exchange.
allow their solutions to a problem to be engaged It is also anormative in this context not to
and modified through additional empirical and listen. Interestingly, hearing is the locus of
theoretical treatments. Certainly, the interaction in this context, not seeing (20). So,
concatenation or assemblage of new material will not listening is misconduct, as would be those
allow each of the contributing pieces to develop practices that chill the aural environment, such as
alternative textures and tones, if not be intruding on others’ exposition, translating a
challenged in new ways. A scholar who attempts person’s words into alternative words, attributing
to prevent such consequences, say, as a peer an exposition to a rationale or condition that has
reviewer who also is the author of an earlier text analytic rather than personalistic origins. For
that is being revised and does not allow another example, saying that this person speaks a certain
author to continue the development of or way because she is of a certain psychological
challenge to the work, could be considered to be type is to reduce the individual’s being to a set of
acting in a way I would label “misconduct.” variables and should be considered unethical.
Constitutive Community/Constitutive
Knowledge. In the fully constitutive context, we Conclusion
have a high degree of organicism as emergent Defining scientific misconduct and discussing
forms and emergent knowledge continually examples and exemplars has become, if not
333
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
popular, then a more broadly based consideration 10. Pepper S. World hypotheses. Berkeley, CA: University
than was evident even a few years ago. However, of California Press; 1942.
in spite of a great deal of commentary, we have 11. Hull D. Science as process: an evolutionary account of
failed to extend our considerations to issues the social and conceptual development of science.
Chicago: University of Chicago Press; 1988.
regarding misconduct within various types of 12. Geison, G L, Holmes, F L, editors. Research schools:
academic communities, especially those that are historical reappraisals. Osiris, 9. Chicago: University of
reflected in the social sciences. It is not enough Chicago Press; 1993
to extend concepts having value in one domain of 13. Shapin S. A social history of truth: civility and science
scholarship and then apply them conveniently to in seventeenth century England. Chicago: University of
others. The social sciences and the humanities Chicago Press; 1994.
reflect alternatives that have meaning for 14. Pickering A. The mangle of practice: time, agency, &
misconduct, and not only the conduct of work. science. Chicago: University of Chicago Press; 1995.
While academic communities have legitimate 15. Silverman R. The impact of electronic publishing on the
academic community. Peek, R P, Newby G B, editors.
interests regarding fabrication, falsification, and Scholarly publishing: the electronic frontier. Cambridge:
plagiarism, there has been no previous attempt to MA: MIT Press; 1995.
go the basics and to consider what might be 16. Silverman R. Comments and replies: academic
anormative for different ways of knowing in conversations. Science Communication 1994; 2: 132-
different settings, not in a methodological, but in 151.
an epistemological way. 17. Detienne M. The masters of truth in archaic Greece.
This paper has attempted to explore New York: Zone; 1996.
“misconduct,” with the explicit understanding 18. Silverman R. Publishing in an immature field: the case
that the ideas and possibilities discussed here are of ‘higher education.’ Higher Education Research and
Development 1986; 5,2: 123-134.
not presented as answers or solutions but as 19. Caws, M A. The art of interference: stressed readings in
heuristic tools to carry the initial discussion. It is verbal and visual texts. Princeton: Princeton University
likely, even hoped, that what has been noted here Press; 1989.
will be revised in the continuing dialogue that 20. Levin D M. The listening self: personal growth, social
transcends these notions and goes to deeper and change, and the closure of metaphysics. London:
more critical hearts of the matter. And for that Routledge; 1989.
conversation to be based on research would allow
for such engagement.

Bibliography
1. Broad W, Wade N. Betrayers of the truth: betrayers in
the halls of science. New York: Simon & Schuster;
1982.
2. Journal of Higher Education. Special issue on research
misconduct. 1994; 65, #3.
3. Alcoff L, Potter E, editors. Feminist epistemologies.
New York: Routledge; 1993.
4. Mitroff I, Kilmann R. Methodological approaches to
social science. San Francisco: Jossey-Bass; 1978.
5. Commission on Research Integrity. Integrity and
misconduct in science. Washington: U.S. Dept of Health
and Human Services: 1995.
6. Code L. What can she know?: Feminist theory and the
construction of knowledge. Ithaca, N.Y.: Cornell
University Press; 1991.
7. Fuchs S. The professional quest for truth: a social theory
of science and knowledge. Albany, N.Y.: State
University of New York Press; 1992.
8. Sayre K M. Belief and knowledge: mapping the
cognitive landscape. Lanham, MD.: Rowman &
Littlefield Publishers; 1997.
9. McKeon R. On knowing: the natural sciences. Chicago:
University of Chicago Press; 1994.

334
 Keyword Index

A Course 237
Curriculum 189, 231
Academic culture 41
Academic environments 321 D
Academic misconduct 49
Academic policies 127 Deception 113
Academic-industry ties 127 Defining Issues Test 195
Adversarial epistemology 299 Deviant behavior 321
Affective Learning 117 Digits 269
Alcoholism 113 DIT 195
Animal 231 Dynamics of competition and cooperation 299
Animal research 231 E
Anthrax 93
Assessment 179 Efficiency 151
Authorship disputes 163 Engineering ethics 223
Autoradiograms 261 Environmental science 167
Epistemic model 299
B Equity theory 143
Biology 27 Erratum 291
Biomedical research 299 Ethical behavior 73
Blots 261 Ethical conduct 209
Brazil 99 Ethical evaluation 159
Buyer’s market 315 Ethics 105, 189, 237, 255, 315
Ethics education 195
C Ethics in science 223
Ethics of science 299
Case studies 247 Ethics training instrument 247
Cheating 27, 49 Ethnicity 163
Chemistry 27 Evaluation 195
Citation 291 Evaluation research 321
Code of ethics 65
Code of protection 167 F
Cognition 105
Collective agency 315 Faculty development 41
Competence 255 G
Computer-based instruction 247
Conduct guidelines 57 Gender 163
Confidentiality 105 Genetic screening 113
Conflict of interest 127 Graduate education 19
Cooperative epistemology 299 Graduate socialization 41
Proceedings: Investigating Research Integrity (2001) ––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Graduate students 19 Political and military intelligence 299
Gulf War 93 Preservice Training 117
Prevention 143
H Problem-Based Learning 117
Health 105 Problem-based learning curriculum 209
Human subjects 305 Professional ethics 179
Professional identity 41
I Publication 291

Image processing 261 R

Informed consent 93, 105
institutional review boards 65 Rationalization techniques 143
Instruction 241 RCR 41, 209, 223, 231, 237, 241
Integrity 73, 237 RCR curriculum development 203
Interactive 237 RCR education and training 215
Internet 241 Reflection-in-action skills 209
Investigational drugs 93 Reflection-in-experimentation 209
Research 237
J Research ethics 41, 57, 179
research ethics 241
Journals 255 research funding 315
Research integrity 99, 127, 315
L Research integrity education programs 203
Learning assessment 223 Research Misconduct 151
Research misconduct 143, 255, 269
M Research training 41
Researcher ethics 57
Manipulation and falsification 261 Responsible Conduct of Research 117, 241
Market demand 315 Responsible conduct of research 143, 195, 231
Medical education 189 Responsible conduct of research education 203, 223
Medical ethics 189
Medical schools 189 S
Mental health 105
Methodology 321 Scholarship norms 327
Military 93 Science careers 321
Misconduct 159 Science policy 223
Moral advisor 247 Scientific divorce 159
Moral development 179 Scientific integrity 65, 73, 223
Moral hazards in scientific enterprises 299 Scientific misconduct 65, 143, 163, 261, 321
Motivation 27 Self-directed learning 247
Situational influences 73
N Social and human sciences 99
Social science misconduct 327
Norms 19 Social work 65
O Statistical forensics 269
Statistics 255
Organizational deviance 305 Suppression bias 167
Organizational influences 73
T
P
Teaching assessment 223
Paraphrasing 281 Teaching research ethics 99
Perceived misconduct 35 Terminal Digits 269
Persons with Disabilities 117 Training 237
Pilot assessment 215 Training grants 215
Plagiarism 281 Type of Injury 151
336
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– Index

U
Undergraduate education 49
Undergraduate research 27
Uniform distribution 269

W
Waiver 93
Whistleblowers 167
Work environment 247
World Wide Web 241
Writing guides 281

337