INTERNATIONAL JOURNAL OF SELECTION AND ASSESSMENT VOLUME 11 NUMBER 1 MARCH 2003

Investigating the Dimensionality of

Counterproductive Work Behavior
Melissa L. Gruys* and Paul R. Sackett
Washington State University University of Minnesota

The study investigated the dimensionality of counterproductive work behavior (CWB)

by examining the relationships between various counterproductive behaviors. Utilizing
a university alumni sample (N = 343), data was collected through both self-report and
direct judgments of the likelihood of co-occurrence. Eleven categories of CWB were
examined: (1) Theft and Related Behavior; (2) Destruction of Property; (3) Misuse of
Information; (4) Misuse of Time and Resources; (5) Unsafe Behavior; (6) Poor
Attendance; (7) Poor Quality Work; (8) Alcohol Use; (9) Drug Use; (10) Inappropriate
Verbal Actions; and (11) Inappropriate Physical Actions. CWB items and categories
were generally positively related. Multidimensional scaling analysis suggests that the
CWB categories vary on two dimensions: an Interpersonal-Organizational dimension
and a Task Relevance dimension.

Introduction CWB, the focus is on the behavior itself rather than on

the results or consequences of the behavior (e.g., the

O ne of the central themes within the field of

Industrial/Organizational Psychology is mapping
the relationship between individual characteristics (e.g.,
harm which is done). Only intentional behaviors are
included in this definition. While accidental actions may
cause harm, they are not the focus of the study. The
skills, abilities, personality traits) and workplace definition encompasses behavior that is targeted at both
behavior. Understanding these relationships permits individuals and at the organization, as both types of
organizations to select individuals with characteristics actions can have severe consequences on the
linked to desired workplace behaviors. Historically, organization. Behavior of organizational members is
personnel selection research has focused primarily on included but the behavior of outsiders (e.g., clients.
identifying candidates likely to engage in desirable work former employees) is not included. While outsider actions
behavior (e.g., those likely to produce work of high may be a concern, the actions of current employees are of
quality, or to produce work at a high rate). More recently, foremost concern to any organization.
increased attention has been paid to an expanded array of The domain of counterproductive work behaviors is
job behaviors, including the broad domain of undesirable, quite broad. Prior to the early 1980s, there was a
counterproductive behaviors (Campbell 1990). considerable amount of research on individual
Robinson and Greenberg (1998) reviewed a wide counterproductive behaviors in the workplace, such as
variety of definitions of counterproductive behavior that employee theft, pilferage, sabotage, slow and sloppy
are found in the literature. More recently, Sackett and performance, tardiness and absenteeism (e.g., Altheide,
DeVore (2001) provide another thorough review. Based Adler, Adler, and Altheide, 1978; Bensman and Gerver,
on their treatment, we use the following definition of 1963; Cressey, 1953; Gouldner, 1954; Henry, 1978a,
counterproductive work behavior (CWB) in this 1978b; Horning, 1970; Mars, 1973; Robin, 1969, 1970;
research: any intentional behavior on the part of an Roy, 1952, 1959; Taylor and Walton, 1971). However,
organization member viewed by the organization as due to the lack of an accepted framework or theory for
contrary to its legitimate interests. With this definition of looking at such behaviors, these studies were each seen
as an investigation into a specific type of behavior rather
than as an effort to explore CWB in a more general
* Address for correspondence: Melissa L. Gruys, Department of
Management and Decision Sciences, Washington State University, sense.
14204 NE Salmon Creek Ave., Vancouver WA 98686. E-mail: The key question this article addresses is `what is the
[email protected] underlying structure of CWB?' The article looks at the

Blackwell Publishing Ltd 2003, 9600 Garsington Road, Oxford OX4 2DQ, UK and
30 350 Main Street, Malden, MA 02148, USA.
DIMENSIONALITY OF COUNTERPRODUCTIVE WORK BEHAVIOR 31

covariance of counterproductive behaviors in order to (organizational minor), personal aggression (inter-

examine the structure of such behavior. personal serious, including behaviors such as harassment
Two important streams of relevant research are and theft from co-workers), and political deviance
outlined here. The first stream involves attempts to use (interpersonal minor, including behaviors such as
psychological tests to predict employee theft. Reviews by favoritism, gossip, and blaming others for one's mistakes).
Sackett and colleagues (Sackett and Decker, 1979; Sackett More recently, Bennett and Robinson (2000) have
and Harris, 1984; Sackett, Burris, and Callahan, 1989; developed a measure of workplace deviance and
Sackett and Wanek, 1996) detail developments in this presented construct validity evidence. This study
area. The critical point is that integrity tests initially supported the Robinson and Bennett (1995) distinction
developed to predict theft have over time been found to between production and property deviance. However,
predict a variety of counterproductive behaviors other the authors recommend dropping the serious to minor
than theft, including absenteeism, violation of safety dimension, as this dimension represents more of a
rules, and drug and alcohol use on the job. However, this quantitative distinction than a qualitative one. A recent
work has been hindered by the lack of a clear picture of review by Sackett and DeVore (2001) argues for a
the underlying structure of CWB. hierarchical model of counterproductive behavior in
The second stream involves attempts at the develop- which there is a single general factor under which there
ment of taxonomies of counterproductive work beha- are two more specific factors represented by property and
viors. A seminal body of work examining a broad range production deviance. Within the two more specific
of counterproductive behaviors is that of Hollinger and factors would fall even more specific types of behavior
Clark (1982, 1983a, 1983b) who developed a broad list of (e.g., absenteeism, theft, etc.). Thus, research in the area
counterproductive behaviors, provided a conceptual supports that counterproductive workplace behavior
framework for interrelating those behaviors, and likely has a general factor with more specific factors
collected self-report data from large employee samples representing organizational deviance and interpersonal
in three industries. They proposed that deviance beneath that.
counterproductive behaviors could be grouped into two While the current literature provides some insight into
broad categories. The first category is `property the structure of CWB, there remain questions regarding
deviance', involving misuse of employer assets. Examples the dimensionality of such behavior. For example, within
include theft, property damage, and misuse of discount either organizational deviance or interpersonal deviance,
privileges. The second category is `production deviance', are there various behaviors that are exhibited together
involving violating norms about how work is to be more frequently? How do the behaviors co-vary? One
accomplished. This includes not being on the job as approach to examining the dimensionality of counter-
scheduled (absence, tardiness, long breaks) and behaviors productive behavior is multidimensional scaling, as
that detract from production when on the job (drug and conducted by Robinson and Bennett (1995). A different
alcohol use, intentional slow or sloppy work). Empirical sample and a different approach to ratings would be a
work by Hollinger and Clark (1982, 1983a) has good test of the structure of counterproductive behavior.
supported the distinction between property and It is critical to note that the categorization scheme put
production deviance. forth by Robinson and Bennett (1995) is based on
Robinson and Bennett (1995) noted that the set of workers' perceptions of the similarity of pairs of
behaviors examined by Hollinger and Clark did not behaviors. When asked to judge the similarity of pairs
include interpersonal counterproductive behaviors, such of behaviors. respondents were not constrained as to the
as sexual harassment, and set out to expand upon the basis for their similarity judgments. The purpose of the
Hollinger and Clark framework. They had workers multidimensional scaling technique is to identify
generate a large number of critical incidents of dimensions underlying similarity judgments, and their
counterproductive behavior, obtained ratings of the study findings indicate that, in the aggregate, respondents
similarity between pairs of behaviors, and subjected the used the two dimensions of organization as target vs.
resulting pairwise similarity matrix to multidimensional other person as target and minor offense vs. serious
scaling. They obtained a two-dimensional solution, with offense as the basis for their similarity judgments. While
one dimension differentiating behavior toward the it is useful to know what dimensions underlie perceptions
organization (Hollinger and Clark's production and of similarity between counterproductive behaviors, we
property deviance) from interpersonal behavior toward posit here that a key issue for understanding inter-
other organizational members (e.g., harassment, gossip, relationships among various forms of counterproductive
verbal abuse), and the other dimension representing a behavior is the covariance of occurrence among these
continuum from minor to serious offenses. Arraying behaviors. The question is whether individuals who
behaviors in this two-dimensional space, Robinson and engage in one form of counterproductive behavior are
Bennett labeled the resulting four quadrants as property also likely to engage in others. It is possible that the
deviance (organizational serious), production deviance pattern of co-occurrence among counterproductive

Blackwell Publishing Ltd 2003 Volume 11 Number 1 March 2003

32 MELISSA L. GRUYS AND PAUL R. SACKETT

behaviors is quite different from the pattern of perceived counterproductive behavior. Thus, the study involved
similarity emerging from the Robinson and Bennett two types of data: self-ratings and co-occurrence ratings,
perceptual similarity task. For example, theft from a co- and can be viewed as two parallel attempts to examine
worker and verbal abuse of a customer are located close the covariance among counterproductive behaviors. We
together in the multidimensional space emerging from present these two types of data as Study 1 and Study 2
Robinson and Bennett's analysis, in that both are serious respectively.
offenses that target another person. However, the two
behaviors may differ on a variety of other dimensions
(e.g., a public vs. private dimension, a planned vs. Study 1: Self-Ratings of Counterproductive
unplanned dimension), making the rate of co-occurrence Behavior
of the two an open question.
The co-occurrence of counterproductive work Sample
behaviors is of considerable interest in that an under- Alumni of a small Midwestern undergraduate liberal arts
standing of the interrelationships among these behaviors institution were surveyed. This sample allowed for
can help guide subsequent work on antecedents of inclusion of individuals from a wide variety of
counterproductive behaviors and on interventions aimed occupations and industries who were employed in a
at reducing the occurrence of counterproductive wide variety of organizations across the United States.
behaviors. A finding of high rates of co-occurrence Surveys were mailed to 1,102 alumni and were returned
among particular sets of counterproductive behaviors by 363 for a response rate of 33%. Ten of the surveys
may suggest common antecedents for these behaviors arrived after the deadline and could not be included in
and the possibility that an intervention aimed at one may the analyses and ten additional surveys were eliminated
also influence others in the set. If, for example, all due to incomplete data. The resulting sample size for the
counterproductive behaviors co-varied highly, resulting study was 343.
in a single general factor, a common set of antecedents Some 94% of the sample were White, 56% were
for all of these behaviors might be implicated. At the female, 73% were married, and the mean age of
other extreme, a finding of low covariance might suggest participants was 40 years (SD 10.15). The sample
the need for separate research programs for each was one in which the education level of participants was
counterproductive behavior. It should be noted, however, extremely high, with 59% of the sample having education
that high intercorrelations between behaviors do not beyond an undergraduate degree. Some 94% of the
necessarily mean that variance associated with specific sample were employed and 73% were employed full-
behaviors is unimportant. time.1 Some 86% of the participants were employed in
How might one get insight into the rate of co- managerial, technical, and professional positions. The
occurrence of counterproductive behaviors on the part of sample had mean tenure in current job and job
employees? The ideal would be the objective measure- experience of 8.37 years (SD 8.04) and 15.5 years
ment of each form of counterproductive behavior. We (SD 9.82), respectively, and a mean income of $55,400
acknowledge what has been viewed as the Achilles heel (SD 57.39).
of counterproductive behavior research, namely, that The survey described below was pilot-tested on 115
while some forms of counterproductive behavior are undergraduate students enrolled in management classes
public (e.g., absence), many are acts by employees who at a large Midwestern university. The mean age of
do not wish to be detected (e.g., theft, sabotage, harass- participants was 21.7 years (with a range of 19 to 43);
ment). In the face of the difficulties of direct observation, 55% were female, and 87% were White. Some 79% of
data on the covariance of counterproductive behaviors participants were employed. For those employed, mean
might come from various sources, such as self-reports of work hours per week was 21.5 (SD 9.46), and mean
the rate of occurrence by employees themselves, tenure in present job was 18 months (SD 18.37).
judgments by others (e.g., supervisors) of the rate of
occurrence, and direct judgments about the rate of co-
occurrence of counterproductive behaviors by employees.
Survey Construction
As none of these strategies is clearly ideal, we make use
of dual strategies in the current research. First, we obtain A list of over 250 counterproductive work behaviors was
self-reports of the likelihood of engaging in various compiled from the literature for use in the study. Keyword
counterproductive behaviors. Examining correlations searches were conducted using major databases (e.g.,
among the various behaviors will shed light on the PSYCInfo, Academic Universe, etc.) to identify relevant
dimensionality issue. Second, we obtain direct judgments literature that may contain relevant behaviors. Literature
about the likelihood of co-occurrence of pairs of in the areas of psychology, management, business, and
behavior categories, and transform these judgments into sociology was examined. Items were obtained from 15
a matrix of similarities between different categories of different articles.2 Redundant items were removed, and

International Journal of Selection and Assessment Blackwell Publishing Ltd 2003

DIMENSIONALITY OF COUNTERPRODUCTIVE WORK BEHAVIOR 33

ultimately 66 behaviors were included in the study. The 1975; Fishbein, 1980), which posits intentions as the most
behavior list was then sorted into 11 separate categories of direct precursor to behavior.
CWB based on the similarity of content. The following 11 In addition, the survey also included a section that
categories of behaviors were formed: (1) Theft and asked participants to provide demographic and
Related Behavior; (2) Destruction of Property; (3) Misuse background information. The variables collected were:
of Information; (4) Misuse of Time and Resources; (5) age, sex, race, education, employment status (employed
Unsafe Behavior; (6) Poor Attendance; (7) Poor Quality or not), occupation and industry, tenure, and amount of
Work; (8) Alcohol Use; (9) Drug Use; (10) Inappropriate full-time work experience.
Verbal Actions; and (11) Inappropriate Physical Actions.
Table 1 presents the CWB categories and the behaviors in
each category. The initial sort of the behaviors into the Analysis
categories was conducted by the first author. In a group
session eight doctoral students familiar with the research The central analysis for Study 1 involved combining the
project and its objectives independently sorted the items 66 items into the 11 conceptually derived composites,
into the categories. There were virtually no discrepancies computing the intercorrelations among the 11 com-
between judges, and thus no changes were made to the posites, and conducting a principal components analysis
initial item sort. of this correlation matrix. Exploratory analysis was
Since the study aimed to examine the structure of done. as there was no strong a priori theory as to the
CWB, the survey included a wide range of items. The dimensionality of these 11 behavior dimensions.
Bennett and Robinson (2000) measure of workplace Prior to these analyses, the internal consistency
deviance was not used because the measure did not reliability of the 11 proposed composites was examined,
represent a broad enough range of behaviors. Bennett and confirmatory factor analysis was conducted to
and Robinson acknowledge that in an effort to develop a examine the proposed 11 dimensions. Absent strong
reliable and valid scale, they have included only the most conceptual alternative hypotheses, we tested a single
common forms of employee deviance. The 11 categories factor model and the hypothesized 11- factor model.
of behavior used in the current study include the whole It is acknowledged here that there is a potential for
range of behavior addressed by Bennett and Robinson confusion in Study 1 regarding the use of factor-analytic
(2000). For many of the Bennett and Robinson items, on procedures for two very different purposes. First, factor-
the current survey, there were several more detailed items analytic procedures were used at the item level to test the
related to that particular behavior. viability of combining the 66 items into the 11 categories
Participants were asked to rate whether they would of behavior. Confirmatory factor analysis was used to
engage in each of the counterproductive behaviors on a create and justify the use of the behavior categories.
seven-point scale with 1 being anchored with `No Matter Second, exploratory factor-analytic procedures were used
What the Circumstances, I Would Not Engage in the at the category level to carry out the main purpose of the
Behavior' and 7 being anchored with `In a Wide Variety study, to examine the dimensionality of the 11 behavior
of Circumstances, I Would Engage in the Behavior.' categories. This results in an unusual data analytic
Individuals were asked to respond according to the pattern, namely, that we use confirmatory factor analysis
assumption that they would have the opportunity in their in the first stage of our analysis and exploratory factor
workplace situation to engage in all of the listed analysis in the second. This is intentional: the first stage
behaviors. This type of rating avoids a situation in of the project involved determining whether a set of items
which an employee may not have the opportunity to conformed to a hypothesized 11-factor solution (hence a
engage in one of the behaviors being assessed. Thus, all confirmatory strategy) while the second stage involved
items are equally applicable to participants regardless of examining the structure of the 11 categories (for which
the job in which they are employed. In addition, the we did not have strong a priori hypotheses and thus used
wording of the anchors on the rating scale attempted to an exploratory strategy).
take into account the circumstances under which an
individual would engage in various behaviors. The
instructions for the survey prompted participants to Results
consider various circumstances when rating the items
(e.g., working conditions, penalties or punishments for Table 1 presents means and standard deviations for the
wrongdoing, unfair treatment, level of job satisfaction). items for each CWB category. Mean responses ranged
Thus the study did not ask for reports of behavior, but from 1 to 4.7 on the seven-point scale and most items had
rather of behavioral intentions. There is extensive means of less than 3. These means indicate that
research documenting the very strong relationship participants reported that they are not very likely to
between intentions and behavior, much of it linked to engage in most of the behaviors included on the survey, a
the Theory of Reasoned Action (Fishbein and Ajzen, result not surprising given the nature of the behaviors

Blackwell Publishing Ltd 2003 Volume 11 Number 1 March 2003

34 MELISSA L. GRUYS AND PAUL R. SACKETT

Table 1: Descriptive statistics for counterproductive work behavior categories and items

Category and Items Number Alpha Composite Mean Std. Dev.

of items Estimate Mean

Theft and Related Behavior 10 .77 1.71

Help another person or advise them how to take 1.15 0.52
company property or merchandise.
Take cash or property belonging to the company. 1.66 1.12
Misuse business expense account. 1.55 0.99
Take cash or property belonging to a co-worker. 1.06 0.30
Take office supplies from the company. 2.76 1.65
Take petty cash from the company. 1.15 0.55
Take cash or property belonging to a customer. 1.06 0.25
Give away goods or services for free. 2.97 1.82
Provide goods or services at less than the price 1.89 1.39
established by the company.
Misuse employee discount privileges. 1.82 1.23

Destruction of Property 4 .66 1.11

Deface, damage, or destroy property, belonging 1.10 0.39
to a co-worker.
Deface, damage, or destroy property, belonging 1.07 0.29
to a customer.
Deface, damage, or destroy property, equipment, 1.17 0.47
or product belonging to the company.
Deliberately sabotage the production of product 1.09 0.44
in the company.

Misuse of Information 5 .71 1.57

Destroy or falsify company records or documents. 1.34 0.68
Discuss confidential matters with unauthorized 2.06 1.29
personnel within or outside the organization.
Intentionally fail to give a supervisor or co-worker 1.57 0.93
necessary information.
Provide the organization with false information to 1.20 0.51
obtain a job (i.e., regarding education or
experience).
Lie to employer or supervisor to cover up a mistake. 2.72 1.49

Misuse of Time and Resources 13 .90 2.81

Conduct personal business during work time. 3.41 1.80
Spend time on the internet for reasons not 3.47 1.99
related to work.
Take a long lunch or coffee break without approval. 3.12 1.98
Waste time on the job. 2.89 1.49
Waste company resources. 1.93 1.07
Use company resources you aren't authorized to use. 2.08 1.26
Make personal long distance calls at work. 2.86 1.85
Mail personal packages at work. 2.21 1.70
Make personal photocopies at work. 4.28 1.95
Use email for personal purposes. 4.66 2.17
Play computer games during work time. 2.14 1.58
Alter time card to get paid for more hours than 1.41 0.88
you worked.
Work unnecessary overtime. 2.02 1.57

Unsafe Behavior 4 .71 1.97

Endanger yourself by not following safety 1.94 1.15
procedures.

International Journal of Selection and Assessment Blackwell Publishing Ltd 2003

DIMENSIONALITY OF COUNTERPRODUCTIVE WORK BEHAVIOR 35

Endanger coworkers by not following safety 1.34 0.65

procedures.
Endanger customers by not following safety 1.31 0.62
procedures.
Fail to read the manual outlining safety procedures. 3.28 1.66

Poor Attendance 5 .77 2.06

Be absent from work without a legitimate excuse. 1.87 1.30
Intentionally come to work late. 1.86 1.32
Use sick leave when not really sick. 2.70 1.80
Leave work early without permission. 2.59 1.77
Miss work without calling in. 1.25 0.69

Poor Quality Work 3 .86 1.37

Intentionally perform your job below 1.34 0.76
acceptable standards.
Intentionally do work badly or incorrectly. 1.24 0.63
Intentionally do slow or sloppy work. 1.51 0.88

Alcohol Use 3 .59 1.35

Come to work under the influence of alcohol. 1.11 0.49
Have your performance affected due to a hangover 1.48 0.91
from alcohol.
Engage in alcohol consumption on the job. 1.44 1.06

Drug Use 4 .71 1.04

Engage in drug use on the job. 1.03 0.16
Come to work under the influence of drugs. 1.04 0.22
Possess or sell drugs on company property. 1.02 0.15
Have your performance affected due to a hangover 1.07 0.33
from drugs.

Inappropriate Verbal Actions 8 .82 1.83

Argue or fight with a co-worker. 2.89 1.41
Yell or shout on the job. 1.91 1.06
Verbally abuse a customer. 1.29 0.65
Verbally abuse a co-worker. 1.41 0.84
Verbally abuse a supervisor. 1.27 0.72
Use sexually explicit language in the workplace. 1.84 1.28
Argue or fight with a supervisor. 2.28 1.38
Argue or fight with a customer. 1.65 0.96

Inappropriate Physical Actions 7 .82 1.08

Physically attack (e.g., pushing, shoving, hitting) 1.11 0.33
a co-worker.
Physically attack (e.g., pushing, shoving, hitting) 1.04 0.20
a customer.
Physically attack (e.g., pushing, shoving, hitting) 1.06 0.27
a supervisor.
Make unwanted sexual advances toward a 1.11 0.49
subordinate.
Make unwanted sexual advances toward a 1.07 0.36
supervisor.
Make unwanted sexual advances toward a 1.10 0.44
co-worker.
Make unwanted sexual advances toward a 1.07 0.28
customer.

Note. The composite means are reported as a mean item for comparability.

Blackwell Publishing Ltd 2003 Volume 11 Number 1 March 2003

36 MELISSA L. GRUYS AND PAUL R. SACKETT

being rated. However, there were some behaviors which Time and Resources; Unsafe Behavior, and Poor
sizable portions of the sample endorsed as being Attendance. The lowest means for the CWB categories
behaviors in which they would be quite likely to engage, are those for Destruction of Property, Drug Use, and
such as `Use email for personal purposes' (item mean Inappropriate Physical Actions.
4.66 on a seven-point scale), `Make personal photocopies Most of the background and demographic variables
at work' (mean 4.28), `Use internet for non-work were relatively uncorrelated with the 11 categories of
purposes' (mean 3.47), and `Use sick leave when not CWB, with the exception of age and work experience.
really sick' (mean 2.70). This result is consistent with There were significant negative relationships between age
the findings of Bennett and Robinson (2000) who found and six of the 11 categories. This indicates a trend of
that for many of the counterproductive behavior items older individuals being less likely to engage in these types
surveyed, there was a relatively high participation rate. of counterproductive behavior. Work experience has a
Thus, it is not the case that all behaviors that at least similar negative relationship with the counterproductive
some organizations view as counterproductive are behavior categories. This is not surprising given the
uncommon. correlation of .86 between age and work experience. In
We note that means and standard deviations for all general, these results are not unexpected given the results
individual CWB items were very consistent between the of past research which suggests that older individuals and
student pilot sample and the main study sample. The those with longer tenure are less likely to engage in
correlation between the individual item means for the counterproductive behavior in the workplace (e.g..
two samples was .95 and the correlation between Hollinger and Clark, 1983b).
individual item standard deviations was .92. These The correlations between the CWB categories ranged
findings are striking given how different the two samples from .17 to .71 and are all positive and significant at the
are in terms of age as well as work and life experience. p<.01 level. The average correlation is .43. This is
This suggests that the findings are not specific to the similar to the correlation of .46 (uncorrected) between
alumni sample used here. organizational and interpersonal deviance found by
Table 1 presents the alpha estimates of reliability for Bennett and Robinson (2000).
each of the CWB category composites. The alpha Principal components analysis was conducted on the
estimates of reliability for the categories exceed .7, with categories of behaviors using an oblique rotation. The first
the exception of Destruction of Property and Alcohol Use two eigenvalues were 5.51 and 1.17. Thus a scree plot
which have alpha estimates of .66 and .59, respectively. suggests a 1-factor solution, while an eigenvalue-greater-
As an additional check on the appropriateness of the than-one rule of thumb suggests two factors. We examined
11 categories used here, confirmatory factor analyses both solutions. The 2-factor solution produces two highly
were conducted using LISREL. The first model that was correlated factors (r :50), and the second factor is
run in LISREL was a single factor model that specified clearly methodological rather than substantive: factor 1
that all 66 of the items would load on a single represents dimensions with high means and factor 2
counterproductive behavior factor. This model produced represents dimensions with low means. Thus we focused
an extremely poor fit. The Goodness of Fit (GFI) index on the 1-factor solution, and present the results in Table 3.
was .50, the Adjusted Goodness of Fit (AGFI) was .46, All of the categories load quite highly on the first factor.
the Standardized Root Mean Square Residual (SRMR) Thus, the results suggest a strong common dimension
.093, and the Root Mean Square Error of Approximation underlying CWB. All categories load highly on the first
(RMSEA) was .12. An 11-factor model was run using factor, which may be characterized as general
item parcels of 35 items per parcel (with items being counterproductive work behavior. These results suggest
randomly assigned within category to a parcel). The fit of that as the likelihood of an individual engaging in a
an 11-factor model was substantially better with certain type of CWB increases, the likelihood of that
goodness of fit statistics as follows: the GFI was .83, individual engaging in a wide variety of other types of
the AGFI was .76, the SRMR was .083, and the RMSEA CWB also increases.
was .064. All of the maximum likelihood estimates for
the paths were significant (p<.01). Though the fit of this
model was better, the fit was moderate.
Items making up each of the 11 dimensions were then Study 2: Co-occurrence Ratings of
summed to form composites. Table 2 presents means, Counterproductive Work Behavior
standard deviations, and intercorrelations for the 11
Sample
dimensions and for other demographic variables.
Reliabilities are indicated on the diagonal. Given that The same sample of 343 college alumni used in Study 1
the number of items per CWB category varies, the means was used in Study 2. Both the self-rating task described
for the category composites that are presented are the above and the co-occurrence ratings described here were
mean per item. The highest means are for Misuse of included in the same mail survey.

International Journal of Selection and Assessment Blackwell Publishing Ltd 2003

 Blackwell Publishing Ltd 2003

DIMENSIONALITY OF COUNTERPRODUCTIVE WORK BEHAVIOR

Table 2: Descriptive statistics and intercorrelations for all variables

Variable Mean1 SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1. Theft 1.71 0.64 (.77)

2. Property 1.11 0.29 .47** (.66)
3. Informat 1.57 0.62 .71** .55** (.71)
4. Timereso 2.81 1.13 .74** .43** .71** (.90)
5. Safety 1.97 0.81 .46** .35** .57** .48** (.71)
6. Attend 2.06 1.04 .63** .38** .64** .76** .37** (.77)
7. Workqual 1.37 0.68 .55** .51** .64** .50** .33** .53** (.86)
8. Alcohol 1.35 0.63 .38** .30** .35** .40** .33** .29** .20** (.59)
9. Drugs 1.04 0.17 .27** .44** .29** .21** .17** .27** .38** .28** (.71)
10. Verbal 1.83 0.72 .55** .52** .57** .54** .47** .47** .42** .39** .27** (.82)
11. Physical 1.08 0.24 .42** .59** .42** .36** .24** .31** .26** .29** .38** .43** (.82)
12. Gender 1.44 0.50 .07 .18** .01 .00 .07 .06 .02 .12** .06 .12* .15**
13. Age 39.77 10.15 .15** .05 .18** .30** .08 .24** .12* .04 .09 .08 .11* .24**
14. Married 0.73 0.45 .04 .03 .07 .13* .07 .04 .11* .00 .02 .04 .09 .12* .27**
15. Educ. 17.74 2.05 .02 .06 .14* .07 .07 .06 .03 .01 .03 .09 .09 .12* .12* .11
16. Employed 0.93 0.25 .10 .06 .06 .13* .08 .08 .04 .02 .05 .07 .08 .12* .02 .04 .09
17. Orgtype 0.47 0.50 .08 .08 .07 .15** .00 .07 .06 .07 .05 .04 .08 .11* .00 .02 .56** .11
18. Tenure 8.36 8.09 .08 .00 .09 .15** .10 .10 .06 .04 .04 .03 .03 .26** .57** .22** .07 .11 .02
19. Hrs/Wk. 42.01 12.06 .06 .03 .10 .04 .08 .10 .06 .03 .04 .03 .04 .34** .03 .07 .17** .31** .13* .20**
20. Work 15.50 9.82 .13* .02 .19** .27** .05 .24** .14** .02 .07 .04 .07 .35** .86** .24** .11 .01 .04 .56** .21**
Exp
Volume 11 Number 1 March 2003

21. Income 55.38 57.39 .01 .03 .09 .01 .03 .01 .13* .02 .06 .01 .05 .26** .31** .17** .23** .11 .24** .22** .30** .35**

Notes:
1
For comparability, the means for category composites reflect the mean per item. Alpha estimates of reliability are along the diagonal when available.
Sample sizes range from N 318 to N 343 due to missing data. For variables 1 through 11, the counterproductive behavior categories, Theft Theft and Related Behavior; Property Destruction of Property;
Informat Misuse of Information; Timereso Misuse of Time and Resources; Safety Unsafe Behavior; Attend Poor Attendance; Workqual Poor Quality Work; Alcohol Alcohol Use; Drugs Drug Use; Verbal
Inappropriate Verbal Actions; Physical Inappropriate Physical Actions. Gender was a dichotomous variable which was coded as follows: 0 female and 1 male. Married was a dichotomous variable which was
coded as follows: 1 married and 0 not married. Education represents the total years of education. Employed was a dichotomous variable which was coded as follows: 0 not employed and 1 employed. Orgtype
represents the type of organization employed in and was coded as follows: 0 non-profit and 1 for profit. Tenure and work experience were measured in years. Hrs/Wk. represents the number of hours worked per
week. Income was measured in dollars. * p < :05. ** p < :01.

37
38 MELISSA L. GRUYS AND PAUL R. SACKETT

Table 3: Alumni sample factor loadings for principal could not be used, as the mail survey would not ensure an
components analysis of counterproductive work adequate sample size for any given pair of items.
behavior categories
Analysis
One Factor Solution
Multidimensional scaling analysis using ALSCAL was
Category Factor Loading conducted using the data from the co-occurrence ratings.
Participants rated all possible pairs of CWB categories.
Misuse of Information .87 The co-occurrence ratings were used to directly form a
Theft and Related Behavior .83 similarity matrix for use in the multidimensional scaling
Misuse of Time and Resources .82 analysis.
Poor Attendance .76
Inappropriate Verbal Actions .74
Destruction of Property .71 Results
Poor Quality Work .70
Unsafe Behavior .63 Table 4 presents the similarity matrix. All of the mean
Inappropriate Physical Actions .60 similarities between the pairs of items were in a fairly
Alcohol Use .53 small range, from 3.26 to 5.45. This indicates that, in
Drug Use .48 general, participants view all of the 11 CWB categories as
co-occurring to a certain degree. However, the
Note: N 343. similarities suggest that participants do not view any of
the categories as extremely likely to co-occur, which
would be reflected by mean similarities closer to seven.
Survey Construction As input to the multidimensional scaling program, the
Based on preliminary empirical support for the 11 similarity matrix presented in Table 4 was converted into
categories of CWB from Study 1, the co-occurrence a dissimilarity matrix by subtracting each of the
rating task was performed using these categories. The similarities from a constant. ALSCAL was used to
survey asked participants to rate the co-occurrence of produce solutions for one to five dimensions. A scree
categories of CWB. This task involved the comparison of test was conducted by plotting the stress indexes for all
pairs of categories. The survey described co-occurrence five configurations. The appropriate number of
of behaviors as follows: dimensions can be determined on the basis of the scree
plot and where the stress index levels off. The one-
If the behaviors in the two categories co-occur,
dimensional solution had a stress level of .34. The stress
employees who engage in behaviors in one of the
level dropped to .12 for the two-dimensional solution,
categories would also engage in the behaviors in the
and to .05, .03, and .01 for the three- four- and five-
other category. Further, employees who would not
dimensional solutions, respectively. The scree results
engage in behaviors in one of the categories would
suggest either a two-dimensional or three-dimensional
also not engage in behaviors in the other category in
solution. For the sake of interpretability and parsimony,
the pair. If the behaviors in the two categories do not
a two-dimensional solution was chosen. This solution is
co-occur, then employees who engage in behaviors in
presented in Figure 1.
one of the categories would not engage in behaviors in
Dimension 1 can be interpreted as reflecting an
the other category in the pair.
Interpersonal versus Organizational dimension, or the
Participants were given a description of the behaviors degree to which the behaviors are directed at an
that would be included in each category and were asked individual or at the company. This dimension mirrors
to rate the co-occurrence of the pairs of categories on a the Interpersonal-Organizational dimension proposed by
seven point scale with 1 being anchored with `Very the Robinson and Bennett (1995) typology. The behavior
Unlikely to Co-Occur' and 7 being anchored with `Very categories that fall in the most extreme negative area on
Likely to Co-Occur.' They were asked to make their this dimension contain mainly actions that are aimed at
ratings with regard to employees in general rather than in individuals. Inappropriate physical actions (labeled
reference to a specific organization or workplace. Physical in Figure 1), inappropriate verbal actions
The co-occurrence of the categories was rated rather (labeled Verbal), and behaviors which endanger others
than rating the co-occurrence of the 66 individual due to lack of following safety procedures (labeled Safety)
counterproductive work behaviors used in the self-rating are all Interpersonal acts in that they are aimed at
task in Study 1. With 66 items, 2,145 pairs of items would individuals. The behaviors within the categories of
need to be rated; with 11 categories, the number of pairs Alcohol Use and Drug Use are not primarily aimed at
reduces to a more manageable 55. An incomplete design other individuals and are also not necessarily aimed at the

International Journal of Selection and Assessment Blackwell Publishing Ltd 2003

DIMENSIONALITY OF COUNTERPRODUCTIVE WORK BEHAVIOR 39

Table 4: Similarity matrix for co-occurence data

Category 1 2 3 4 5 6 7 8 9 10

1. Theft and Related Behavior

2. Destruction of Property 4.25
3. Misuse of Information 4.17 4.08
4. Misuse of Time and 4.72 4.41 4.88
Resources
5. Unsafe Behavior 3.54 4.47 3.53 3.51
6. Poor Attendance 4.20 4.04 3.75 4.69 3.52
7. Poor Quality Work 4.20 4.51 4.11 4.65 4.24 5.45
8. Alcohol Use 4.28 4.59 3.94 4.26 4.77 5.24 5.31
9. Drug Use 4.50 4.64 3.97 4.28 4.78 5.26 5.29 5.13
10. Inappropriate Verbal 3.67 4.30 3.63 3.34 3.67 3.45 3.89 5.13 5.13
Actions
11. Inappropriate Physical 3.50 4.36 3.26 3.12 3.72 3.30 3.69 5.02 5.06 4.99
Actions

Note: Ratings were made on a 7-point scale with 7 being the most likely to co-occur.

company, which is reflected by their somewhat neutral relevant to tasks that are performed within the context of
location on this dimension. The behavior categories that a job. In general, employees should attend work (Attend),
fall in the most extreme positive area on this dimension use time and resources appropriately (Timereso), have
contain mainly actions that are aimed at the organization high quality work (Workqual), and not engage in
rather than at individuals and are more impersonal activities that would put themselves or others in danger
actions. Poor Attendance (Attend), Misuse of Time and (Safety, Alcohol and Drugs). The behavior categories in
Resources (Timereso), and Misuse of Information the negative area reflect activities that are not related to
(Informat) are all categories that contain such behaviors. tasks performed within the context of a job. Engaging in
Theft and Related Behavior also falls in the positive area inappropriate physical or verbal actions toward others
for the Interpersonal-Organizational dimension. (Physical and Verbal), theft and related behavior (Theft),
Dimension 2 can be interpreted as reflecting Task misuse of property (Property), and misuse of information
Relevance. All of the behavior categories that fall on the (Informat) can all be viewed as somewhat separate from
positive area of this dimension can be described as being specific work tasks that must be performed within jobs.

Figure 1. Multi-dimensional scaling analysis solution for the co-occurrence rating data

Blackwell Publishing Ltd 2003 Volume 11 Number 1 March 2003

40 MELISSA L. GRUYS AND PAUL R. SACKETT

Discussion international context. Along these lines, Murphy (1993)

mentions several cross-cultural differences in the
This research yields a number of interesting results. interpretations of possible counter-productive behaviors.
Correlational analyses of self-ratings in Study 1 indicate a The two studies both produced a matrix indicating the
general pattern of positive correlations between all CWB interrelationship among the same 11 categories of CWB.
items. This indicates that as the likelihood that an Each study provides insight into the dimensionality of
individual will engage in one type of CWB increases, the CWB in a different way. In the case of the Study 1 self-
likelihood of the individual to engage in another type of ratings, these interrelationships were indexed by
CWB also increases. correlations. In the case of the Study 2 co-occurrence
Empirical support was found in Study 1 for the use of ratings, these interrelationships were indexed by direct
categories of behaviors that were formed on the basis of judgments of the likelihood of co-occurrence, and the
content themes. Factor analyses suggested a moderate fit similarity matrix was subjected to multidimensional
for an 11-factor model that corresponds to the 11 CWB scaling. The two methods produced somewhat differing
categories. Support for the use of the categories is useful pictures of the dimensionality of CWB.
in that future studies can characterize counterproductive The findings from the self-report data in Study 1 reveal
behavior that occurs in the workplace in terms of these a strong association between all behavior items and
content categories. The categories can be used as a means between all behavior categories. The correlations
of organizing extensive lists of behaviors. Further, between the CWB categories were all positive and ranged
exploring the relationships between these categories will from .17 to .71 and the average correlation was .43. This
continue to contribute to the formation of a taxonomy of is similar to the correlation of .46 (uncorrected) between
CWB. This, in turn, will contribute to the formation of organizational and interpersonal deviance found by
an overall theory of CWB. Bennett and Robinson (2000). Results also reveal a
In general, most of the background and demographic strong common underlying factor. This would suggest
variables were relatively uncorrelated with the categories that as the likelihood of an individual engaging in one
of CWB. However, age and work experience were type of CWB increases, the likelihood of that individual
significantly negatively related to the categories. This engaging in other types of CWB also increases. There are,
negative correlation indicates that older individuals are of course, differences in base rates across behaviors, so
less likely to engage in these types of CWB. In general, this is not to suggest that an individual who will engage
this result is not unexpected given the results of past in one of the counterproductive behaviors will also
research that suggests that older individuals are less likely engage in all other counterproductive behaviors.
to engage in counterproductive behavior in the work- The findings from the co-occurrence data in Study 2
place (e.g., Hollinger and Clark, 1983b). Many of the indicate a two-dimensional solution. The first dimension
past studies which found this result were conducted on is an Interpersonal-Organizational dimension that
samples which were considerably younger and had reflects whether the categories include behaviors that
considerably less range in age than the present sample are aimed at individuals or at the company. This
and many were conducted using samples from retail dimension parallels Robinson and Bennett's (1995)
organizations (e.g., fast food restaurants) which employ Interpersonal-Organizational dimension. The second
many young people. Therefore, it is notable that the dimension reflects Task Relevance, or the extent to
negative relationship between age and CWB still seems to which the categories include behaviors that are related to
hold in this sample where the mean age is 40 years old, as work tasks that are carried out on the job. One extreme
well. Work experience has a similar negative relationship on this dimension includes the Misuse of Time and
with the CWB categories. Resources and Poor Work Quality categories, which
The use of a sample of older, well-educated contain behaviors which are directly related to carrying
individuals who hold primarily managerial and technical out work tasks. The other extreme on this dimension
level positions can be viewed as a limitation of the study includes the Inappropriate Physical Actions,
in the sense that this sample may not represent the Inappropriate Verbal Actions, Theft and Related
general working population well. It is encouraging, Behavior, and Destruction of Property categories, which
though, that the results using our pilot sample of very contain behaviors that are not related to work tasks.
different individuals, namely, undergraduate students, This two-dimensional solution differs from the two-
seem to converge for the self-ratings. However, the dimensional solution found by Robinson and Bennett
student sample may also be viewed as unrepresentative of (1995). One dimension, the Interpersonal-Organizational
the working population. Regardless of the representative- dimension found in this study corresponds to the
ness of either sample, it is encouraging to see the results Interpersonal-Organizational dimension found by
using the self-ratings for both samples converge. Also, Robinson and Bennett, but the other does not. Robinson
since the study was conducted using an American sample, and Bennett's second factor was a Minor-Serious
the results might not generalize to the broader dimension, whereas the second dimension in this study

International Journal of Selection and Assessment Blackwell Publishing Ltd 2003

DIMENSIONALITY OF COUNTERPRODUCTIVE WORK BEHAVIOR 41

is Task Relevance. It appears that when asked to make This study represents an important step toward a better
judgments of similarity without specifying further the understanding of the structure or dimensionality of
definition of `similarity', respondents attend to a Minor- CWB. Examining the relationships between various
Serious dimension more strongly than in the case of the counterproductive behaviors will eventually contribute
present study, where similarity is explicitly defined in to more effective prediction and prevention of CWB in
terms of co-occurrence. the workplace.
The difference in study findings between the Study 1
self-report data and the Study 2 co-occurrence data (as
well as the differences from other research findings, such Acknowledgements
as those of Robinson and Bennett) indicate the need for
care in drawing conclusions about the dimensionality of This article is based on the thesis submitted by Melissa L.
CWB. We had hoped that the two methodologies Gruys as part of the requirements for her doctoral degree
employed here would converge on a common set of in Human Resources and Industrial Relations at the
findings. They do to some degree: in both studies what University of Minnesota. An earlier version of this article
emerges is a picture of positive relationships among the was presented at the 16th Annual Meeting of the Society
various categories of CWB. Study 1, though, suggests a for Industrial and Organizational Psychology, April
single common factor, while Study 2 differentiates among 2001, San Diego, CA.
the CWB categories on two dimensions. We note that the We gratefully acknowledge funding for this study by the
two methodologies do ask participants to take differing Department of the United States Navy, Office of Naval
perspectives: Study 1 asks for self-ratings, while Study 2 Research grant N00014-98-1-0872 through the Personnel
asks for participants' perceptions about whether Security Research Dissertation Award Program. The
counterproductive behaviors co-occur among workers interpretations, conclusions, and recommendations,
generally and that perceptions of self and perceptions of however, are those of the authors and do not necessarily
others may differ. Thus, continued careful work with represent the position or policy of the agency.
multiple methods is encouraged to further clarify the
structure of CWB.
A contribution of the study rests in the adoption of a Notes
focus on the co-occurrence of behaviors. Co-occurrence
is used as a tool to examine the dimensionality of CWB. 1. Since t-tests indicated no mean differences on relevant
While the two data collection strategies operationalize variables between participants who were not employed,
those who were employed part-time, and those who were
co-occurrence differently (i.e., as correlations between employed full-time, all were retained in the sample.
self-reports vs as direct judgments of likelihood of co- 2. The articles from which items were obtained are listed with
occurrence), both operationalize `similarity' of behaviors an asterisk in the reference list.
in terms of co-occurrence. In past literature, there had
been a focus on the `similarity' of behaviors rather than
on whether behaviors are likely to co-occur. Also, in References
contrast to many past studies that address only one type
of counterproductive behavior, the current study Altheide, D.L., Adler, P.A., Adler, P. and Altheide, D.A. (1978)
addresses the relationships between various types of The social meanings of employee theft. In J.M. Johnson and
J.D. Douglas (eds), Crime at the Top (pp. 90124).
CWB. Determining the level of co-occurrence of Philadelphia: Lippincott.
behaviors in the workplace is very informative to *Baron, R.A. and Neuman, J.H. (1996) Workplace violence and
organizations. The prediction and prevention of CWB workplace aggression: Evidence on their relative frequency
in organizations will be enhanced as the level of co- and potential causes. Aggressive Behavior, 22, 161173.
occurrence of the behaviors becomes known. For Bennett, R.J. and Robinson, S.L. (2000) Development of a
measure of workplace deviance. Journal of Applied
example, if there is research to suggest that certain Psychology, 85, 349360.
variables are significant predictors of one type of CWB Bensman, J. and Gerver, J. (1963) Crime and punishment in the
(e.g., theft) and that behavior is known to co-occur with factory: The function of deviancy in maintaining the social
another type of CWB (e.g., destruction of property), the system. American Sociological Review, 28, 588598.
predictors of theft may well be useful in predicting Campbell, J.P. (1990) Modeling the performance prediction
problem in industrial and organizational psychology. In
destruction of property also. Am analogous situation M.D. Dunnette and L.M. Hough (eds), Handbook of
exists for the prevention of CWB. Industrial and Organizational Psychology (Vol. 1, pp. 687
Robinson and Bennett (1995), Robinson and 732). Palo Alto, CA: Consulting Psychologists Press.
Greenberg (1998), Sackett and Devore (2001), and others Cressey, D. (1953) Other People's Money: A Study in the Social
have acknowledged the need for the development of a Psychology of Embezzlement. Belmont, CA: Wadsworth.
Fishbein, M. (1980) A theory of reasoned action: Some
comprehensive theory of CWB that includes all of the applications and implications. In H. Howe and M. Page
different types of CWB that can occur in the workplace.

Blackwell Publishing Ltd 2003 Volume 11 Number 1 March 2003

42 MELISSA L. GRUYS AND PAUL R. SACKETT

(eds), Nebraska Symposium on Motivation, 27, 65116. Rise, J., Astrom, A.N. and Sutton, S. (1998) Predicting inten-
Fishbein, M. and Azjen, I. (1975) Belief, Attitude, Intention, and tions and use of dental floss among adolescents: An applica-
Behavior: An Introduction to Theory and Research. tion of the theory of planned behavior. Psychology and
Reading, MA: Addison-Wesley. Health, 13, 223236.
Gouldner, A.W. (1954) Wildcat Strike: A Study in Worker- Robin, G.D. (1969) Employees as offenders. Journal of
Management Relationships. New York: Harper and Row. Research in Crime and Delinquency, 6, 1733.
Henry, S. (1978a) The Hidden Economy: The Context and Robin, G.D. (1970) The corporate and judicial disposition of
Control of Borderline Crime. London: Martin Robertson. employee thieves. In E.O. Simgel and H.L. Ross (eds),
Henry, S. (1978b) Crime at work: The social construction of Crimes against Bureaucracy (pp. 119142). New York: Van
amateur property theft. Sociology, 12, 245263. Nostrand Reinhold.
*Hollinger, R.C. (1986) Acts against the workplace: social *Robinson, S.L. and Bennett, R.J. (1995) A typology of deviant
bonding and employee deviance. Deviant Behavior, 7, 5375. workplace behaviors: A multidimensional scaling study.
*Hollinger, R.C. and Clark, J.P. (1982) Formal and informal Academy of Management Journal, 38, 555572.
social controls of employee deviance. The Sociology Robinson, S.L. and Greenberg, J. (1998) Employees behaving
Quarterly, 23, 333343. badly: Dimensions, determinants, and dilemmas in the study
*Hollinger, R.C. and Clark, J.P. (1983a) Deterrence in the of workplace deviance. In D.M. Rousseau and C. Cooper
workplace: Perceived certainty, perceived severity and (eds), Trends in Organizational Behavior (Vol. 5, pp. 130).
employee theft. Social Forces, 62, 398418. New York: Wiley.
Hollinger, R.C. and Clark, J.P. (1983b) Theft by Employees. *Robinson, S.L. and O'Leary-Kelly, A. (1996) Monkey see,
Lexington, MA: Heath. monkey do: The role of role models in predicting workplace
*Hollinger, R.C., Slora, K.B. and Terris, W. (1992) Deviance in aggression. In J.B. Keys and L.N. Dosier (eds), Academy of
the fast-food restaurant: Correlates of employee theft, Management Best Paper Proceedings (pp. 288292).
altruism, and counterproductivity. Deviant Behavior: An Pleasantville, NY: Academy of Management.
Interdisciplinary Journal, 13, 155184. Roy, D. (1952) Quota restrictions and goldbricking in a
Horning, D.N.M. (1970) Blue collar theft: Conceptions of machine shop. American Journal of Sociology, 57, 427442.
property, attitudes toward pilfering, and work group norms Roy, D. (1959) Banana time: Job satisfaction and informal
in a modern industrial plant. In E.O. Smigel and H.L. Ross interaction. Human Organization, 18, 158168.
(eds), Crimes against Bureaucracy (pp. 4664). New York: *Ruggiero, M., Greenberger, E. and Steinberg, L.D. (1982)
Van Nostrand Reinhold. Occupational deviance among adolescent workers. Youth
*Hunt, S.T. (1996) Generic work behavior: An investigation and Society, 13, 423448.
into the dimensions of entry-level, hourly job performance. Sackett, P.R. and Decker, P.J. (1979) Detection of deception in
Personnel Psychology, 49, 5183. the employment context: A review and critical analysis.
*Jones, J.W. (1980) Attitudinal correlates of employees' Personnel Psychology, 32, 487506.
deviance: Theft, alcohol use, and nonprescribed drug use. Sackett, P.R. and DeVore, C.J. (2001) Counterproductive
Psychological Reports, 47, 7177. behaviors at work. In N. Anderson, D.S. Ones, H.K.
*Mangione, T.W. and Quinn, R.P. (1975) Job satisfaction, Sinangil and V. Viswesvaran (eds), International Handbook
counter-productive behavior and drug use at work. Journal of Work Psychology (Vol. 1, pp. 145164). London: Sage.
of Applied Psychology, 60, 114116. Sackett, P.R. and Harris, M.M. (1984) Honesty testing for
Mars, G. (1973) Chance, punters and the fiddle: Institu- personnel selection: A review and critique. Personnel
tionalized pilferage in a hotel dining room. In M. Warner Psychology, 37, 221245.
(ed.), The Sociology of the Workplace (pp. 200210). New Sackett, P.R. and Wanek, J.E. (1996) New developments in the
York: Halsted Press. use of measures of honesty, integrity, conscientiousness,
Murphy, K.R. (1993) Honesty in the Workplace. Pacific Grove, dependability, trustworthiness, and reliability for personnel
CA: Brooks/Cole. selection. Personnel Psychology, 49, 787829.
Ones, D.S. and Viswesvaran, C. (1998) Integrity testing in Sackett, P.R., Burris, L.R. and Callahan, C. (1989) Integrity
organizations. In R.W. Griffin, A. O'Leary-Kelly and J.M. testing for personnel selection: An update. Personnel
Collins (eds), Dysfunctional Behaviors in Organizations, Psychology, 42, 491529.
Vol. 2. Nonviolent Behaviors in Organizations. Greenwich, *Skarlicki, D.P. and Folger, R. (1997) Retaliation in the
CT: JAI Press. workplace: The roles of distributive, procedural, and inter-
Ones, D.S. and Viswesvaran, C. (2001) Integrity tests and other actional justice. Journal of Applied Psychology, 82, 434443.
criterion-focused occupational personality scales (COPS) *Slora, K.B. (1989) An empirical approach to determining
used in personnel selection. International Journal of employee deviance base rates. Journal of Business and
Selection and Assessment, 9, 3139. Psychology, 4, 2, 199218.
Ones, D.S., Viswesvaran, C. and Schmidt, F.L. (1993) Compre- Taylor, L. and Walton, P. (1971) Industrial sabotage: Motives
hensive meta-analysis of integrity test validities: Findings and meanings. In S. Cohen (ed.), Images of Deviance (pp.
and implications for personnel selection and theories of job 219245). London: Penguin.
performance. Journal of Applied Psychology, 78, 4, 679703. *Terris, W. and Jones, J. (1982) Psychological factors related to
*Raelin, J.A. (1994) Three scales of professional deviance employees' theft in the convenience store industry. Psycho-
within organizations. Journal of Organizational Behavior, logical Reports, 51, 12191238.
15, 483501.

International Journal of Selection and Assessment Blackwell Publishing Ltd 2003