Archives of Scientific Psychology 2016, 4, 10 –31 © 2016 The Author(s)

DOI: http://dx.doi.org/10.1037/arc0000027 2169-3269

Archives of Scientific Psychology

www.apa.org/pubs/journals/arc

The Single-Case Reporting Guideline In BEhavioural Interventions (SCRIBE)

2016: Explanation and Elaboration

Robyn L. Tate Michael Perdices

The Kolling Institute of Medical Research, St Leonards, New Royal North Shore Hospital, St Leonards, New South Wales,
South Wales, Australia, and The University of Sydney Australia, and The University of Sydney

Ulrike Rosenkoetter Skye McDonald

The Kolling Institute of Medical Research, St Leonards, New University of New South Wales
South Wales, Australia, and The University of Sydney

Leanne Togher William Shadish

The University of Sydney University of California, Merced

Robert Horner Thomas Kratochwill

University of Oregon University of Wisconsin—Madison

David H. Barlow Alan Kazdin

Boston University Yale University

Margaret Sampson Larissa Shamseer

Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada Ottawa Hospital Research Institute, Ottawa, Ontario, Canada,
and University of Ottawa

Sunita Vohra (For the SCRIBE Group)

University of Alberta

A B S T R A C T

There is substantial evidence that research studies reported in the scientific literature do not provide adequate
information so that readers know exactly what was done and what was found. This problem has been addressed by
the development of reporting guidelines which tell authors what should be reported and how it should be described.
Many reporting guidelines are now available for different types of research designs. There is no such guideline for
one type of research design commonly used in the behavioral sciences, the single-case experimental design (SCED).
The present study addressed this gap. This report describes the Single-Case Reporting guideline In BEhavioural
interventions (SCRIBE) 2016, which is a set of 26 items that authors need to address when writing about SCED
research for publication in a scientific journal. Each item is described, a rationale for its inclusion is provided, and
examples of adequate reporting taken from the literature are quoted. It is recommended that the SCRIBE 2016 is used
by authors preparing manuscripts describing SCED research for publication, as well as journal reviewers and editors
who are evaluating such manuscripts.

S C I E N T I F I C A B S T R A C T

Single-case experimental design (SCED) studies in the behavioral sciences literature are not only common, but their
proportion has also increased over past decades. Moreover, methodological complexity of SCEDs and sophistication
in the techniques used to analyze SCED data has increased apace. Yet recent reviews of the behavioral sciences
literature have shown that reporting of SCED research is highly variable and often incomplete. Explicit, precise and
 SCRIBE 2016 EXPLANATION AND ELABORATION 11

transparent reporting is crucial not only for critical evaluation of the study methodology and conclusions, but also to
facilitate exact replication of investigations, and ascertain applicability and possible generality of results. Accord-
ingly, we developed the SCRIBE 2016 (Single-Case Reporting guideline In BEhavioural interventions) by a
consensus process by experts in SCED methodology and research in the behavioral sciences, as well as experts in
reporting guideline development. The SCRIBE 2016 Explanation and Elaboration article describes a set of 26 items
to guide and structure the reporting of SCED research. A rationale and minimum reporting standards that stipulate
what needs to be reported are provided for each item. In addition, examples of adequate and clear reporting drawn
from the literature are included for each item. It is recommended that the SCRIBE 2016 Explanation and Elaboration
article is used in conjunction with the complementary SCRIBE 2016 Statement (Tate et al., 2016) by authors
preparing manuscripts for publication and journal reviewers and editors considering manuscripts for publication.

Keywords: single-case design, methodology, reporting guidelines, publication standards

Supplemental materials: http://dx.doi.org/10.1037/arc0000027.supp

Essentially, without publication, the research remains invisible to the consequences of acquired brain impairment, used single-case
world. And yet, too often, reading these articles leaves us unable to methods (Perdices et al., 2006). This result is comparable to
determine exactly how the research was conducted, what was found, Beeson and Robey’s (2006) findings for the specific domain of
how reliable the findings are and how they fit into the wider context aphasiology (41%). Both Hammond and Gast (2010) and Maggin,
of existing knowledge. Many published articles are not fit for pur-
O’Keeffe, and Johnson (2011) demonstrated an accelerating trend
pose.
for an increased number of single-case reports published over
—(Simera, Moher, Hoey, Schulz, & Altman, 2010, p. 35) recent decades. SCEDs are also used in medicine (Gabler, Duan,
Vohra, & Kravitz, 2011), where they are specifically referred to as
Single-case experimental designs (SCEDs) are used frequently N-of-1 trials. This variety of SCED consists of multiple cross-overs
in the behavioral sciences. Shadish and Sullivan (2011) surveyed (or phase changes) using the withdrawal A-B-A-B paradigm in a
the contents of 21 journals in psychology and education for the single participant who serves as his or her own control, often
calendar year 2008 and found that 44% of intervention studies used incorporating randomization and/or blinding (Kravitz, Duan, & the
single-case methods. Similarly, 39% of records archived on the DEcIDE Methods Center N-of-1 Guidance Panel, 2014). Well-
PsycBITE evidence database (www.psycbite.com), representing all designed and conducted SCEDs provide a strong level of evidence,
published nonpharmacological interventions for psychological and in particular the randomized N-of-1 trial provides Level 1

This article was published April 14, 2016.

Robyn L. Tate, John Walsh Centre for Rehabilitation Research, The Kolling Institute of Medical Research, St Leonards, New South Wales, Australia, and
Sydney Medical School Northern, The University of Sydney; Michael Perdices, Department of Neurology, Royal North Shore Hospital, St Leonards, New South
Wales, Australia, and Discipline of Psychiatry, The University of Sydney; Ulrike Rosenkoetter, John Walsh Centre for Rehabilitation Research, The Kolling
Institute of Medical Research, and Sydney Medical School Northern, The University of Sydney; Skye McDonald, School of Psychology, University of New South
Wales; Leanne Togher, Discipline of Speech Pathology, The University of Sydney; William Shadish, School of Social Sciences, Humanities and Arts, University
of California, Merced; Robert Horner, Department of Special Education and Clinical Services, University of Oregon; Thomas Kratochwill, School of Educational
Psychology, University of Wisconsin—Madison; David H. Barlow, Centre for Anxiety Related Disorders, Boston University; Alan Kazdin, Department of
Psychology, Yale University; Margaret Sampson, Library and Media Services, Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada; Larissa Shamseer,
Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, Ontario, Canada, and School of Epidemiology, Public Health and Preventive
Medicine, University of Ottawa; Sunita Vohra, Department of Pediatrics, University of Alberta.
The SCRIBE Group wishes to pay special tribute to our esteemed colleague Professor William Shadish (1949 –2016) who passed away on the eve of publication
of this article. His contribution at all stages of the SCRIBE project was seminal.
Funding for the SCRIBE project was provided by the Lifetime Care and Support Authority of New South Wales, Australia. The funding body was not
involved in the conduct, interpretation or writing of this work. Members of the SCRIBE Group are as follows: Richard Albin (University of Oregon),
Catherine Backman (University of British Columbia), David H. Barlow (Boston University), Jacinta Douglas (La Trobe University), Jonathan J. Evans
(University of Glasgow), David Gast (University of Georgia), Robert Horner (University of Oregon), Alan Kazdin (Yale University), Thomas Kratochwill
(University of Wisconsin-Madison), Rumen Manolov (University of Barcelona), Skye McDonald (University of New South Wales), Geoffrey Mitchell (The
University of Queensland), Lyndsey Nickels (Macquarie University), Jane Nikles (The University of Queensland), Tamara Ownsworth (Griffith University),
Michael Perdices (Royal North Shore Hospital, New South Wales, Australia, and The University of Sydney), Miranda Rose (La Trobe University), Ulrike
Rosenkoetter (The Kolling Institute of Medical Research, St Leonards, New South Wales, Australia, and The University of Sydney), Margaret Sampson
(Children’s Hospital of Eastern Ontario, Ottawa, Ontario, Canada), Christopher H. Schmid (Brown University), William Shadish (University of California,
Merced), Larissa Shamseer (Ottawa Hospital Research Institute, Ottawa, Ontario, Canada, and University of Ottawa), Robyn L. Tate (The Kolling Institute
of Medical Research, St Leonards, New South Wales, Australia, and The University of Sydney), Leanne Togher (The University of Sydney), Sunita Vohra
(University of Alberta), and Barbara Wilson (Oliver Zangwill Centre, Ely, Cambridgeshire, United Kingdom). For further discussion on this topic, please
visit the Archives of Scientific Psychology online public forum at http://arcblog.apa.org.
This work is licensed under a Creative Commons Attribution 3.0 Unported License (http://creativecommons.org/licenses/by/3.0/), which allows anyone
to download, reuse, reprint, modify, distribute, and/or copy this content, so long as the original authors and source are cited and the article’s integrity is
maintained. Copyright for this article is retained by the author(s). Author(s) grant(s) the American Psychological Association a license to publish the article
and identify itself as the original publisher. No permission is required from the authors or the publisher.
Correspondence concerning this article should be addressed to Robyn L. Tate, John Walsh Centre for Rehabilitation Research, The Kolling Institute of
Medical Research, Level 9, Royal North Shore Hospital, St Leonards NSW 2065, Australia. E-mail: [email protected]
12 TATE ET AL.

evidence1 for treatment decision purposes (Guyatt, Jaeschke, & reported on procedural fidelity. Tate et al. (2014) reported on a random
McGinn, 2002; OCEBM, 2011). sample of 35% of reports archived on the PsycBITE database published
Designs using single-case methods can be complex and sophisticated between 1990 and 2010 in the neurological conditions of dementia, stroke
(e.g., a combination of a multiple-baseline with alternating-treatments and traumatic brain injury that used a single participant (n ⫽ 253). Only
design, this particular method comprising 10% of designs in the Shadish 14% reported using an assessor who was independent of the therapist,
and Sullivan (2011) survey). Recently, a class of single-case designs has 54% reported on interrater reliability of the dependent variable, and 62%
been proposed that involves a variety of randomization procedures and provided a session-by-session data record in graphed or tabular format.
can increase the internal validity of these designs (Kratochwill & Levin, Smith (2012) systematically reviewed the psychology and education
2010), as well as provide options for data analysis (see Kratochwill & literature between 2000 and 2010 specifically to identify SCEDs, with
Levin, 2014). The range of single-case designs, however, is a potential 409 reports meeting eligibility criteria. Twenty-two percent of studies did
source of misunderstanding because not all reports in the literature that not report baseline data and 52% did not report either statistical or visual
study a single participant also apply single-case methodology, as de- analyses of the data. These data from different populations suggest that
fined.2 Figure 1 depicts the common designs using a single-participant as problems with the conduct and/or reporting of fundamental elements of
reported in the literature. Components of the figure and their interrela- single-case research in the behavioral sciences are common and highlight
tionships are described in the companion Single-Case Reporting guide- the need for a reporting guideline.
line In BEhavioural interventions (SCRIBE) 2016 Statement article (Tate Reporting guidelines in the CONSORT tradition (see www
et al., 2016). In brief, surveys of the literature (e.g., Perdices & Tate, .equator-network.org) improve the clarity and transparency of reporting
2009; Shadish & Sullivan, 2011) have identified multiple designs using a of randomized controlled trials (RCT) published in journals that
single participant. The figure presents nine prototypical designs, but not endorse them. A systematic review based on evaluations of more
all of these use single-case methodology, as defined (i.e., containing than 16,000 RCTs found that 25 of 27 CONSORT-related items
multiple phases during each of which the dependent variable is measured were more completely reported in journals that endorse CONSORT
repeatedly; see Footnote 2). In particular, the three designs below the compared to those that do not, with five items being statistically
solid horizontal line (B-phase training study, pre–post intervention eval- significant (p ⫽ .01; Turner et al., 2012). It is expected that the
uations alone, and case description) do not meet these criteria and they are development and implementation of similarly structured guidelines
not SCEDs. to cover common research designs used in SCEDs will assist (a)
The SCRIBE 2016 applies to those designs above the solid horizontal researchers to report on the requisite items that foster complete-
line, which all use single-case methodology, but differ fundamentally in ness, clarity, transparency and accuracy of reporting and (b) read-
terms of their structure. The four prototypical designs above the dotted ers to know exactly what was done and what was found. The
horizontal line comprise the withdrawal/reversal, multiple-baseline, present report, referred to as the SCRIBE 2016 Explanation and
alternating/simultaneous-treatments, and changing-criterion designs. Re- Elaboration article, provides description of and rationale for 26
porting requirements differ among the designs, and are described in the reporting items, along with examples of adequate reporting from
SCRIBE 2016 Item 5. The medical N-of-1 trial falls within the with- the published literature. A separate SCRIBE 2016 Statement (Tate
drawal/reversal paradigm, and a separate reporting guideline is available et al., 2016) describes the methodology underlying development of
for that design (see Shamseer et al., 2015 and Vohra et al., 2015 for the the SCRIBE 2016.
CONSORT Extension for N-of-1 Trials [CENT 2015]). Randomization The genesis of the SCRIBE 2016 derives directly from the
of elements in all of the foregoing designs is feasible (see Kratochwill & foundation work and development of the CENT 2015 (Shamseer et
Levin, 2010), albeit not common practice, and is covered in the SCRIBE al., 2015; Vohra et al., 2015). Frequently, the medical interventions
2016 Item 8. being evaluated in N-of-1 trials use pharmacological or nonphar-
Each of the designs also varies in terms of methodological rigor and macological substances that are injected, ingested, inhaled, or
even experimental designs above the dotted horizontal line may not topically applied. The design for these types of interventions
meet design standards, such as those of Horner et al. (2005) and sometimes involves run-in/wash-out periods, which are conse-
Kratochwill et al. (2013). The reason that the biphasic A-B design is quently highly relevant to adequate reporting of such trials and
separated from the other designs above the dotted horizontal line is specific items are required for the purpose of reporting. By con-
because of its poor control of threats to internal validity. It is for this trast, the SCRIBE 2016 is intended to apply to the broader variety
reason that the A-B design, although using single-case methodology, of experimental single-case interventions used in the behavioral
is regarded as quasiexperimental (Barlow, Nock, & Hersen, 2009). sciences, including, but not restricted to, health conditions. Behav-
Scientific quality for both internal and external validity can be mea-
sured with methodological quality rating scales designed for single- 1
Levels of evidence refer to the hierarchy of strength or credibility of
case methodologies, such as those described in Maggin et al. (2014) evidence that different research designs yield. The hierarchy is frequently used
and Tate et al. (2013b). in medicine to critically evaluate the available evidence for different clinical
The behavioral sciences literature is highly variable with respect to the questions (e.g., interventions, harms, diagnosis, prognosis). There is some
variation among different classification systems (see, e.g., websites of the
adequate conduct and complete reporting of single-case research. The American Academy of Neurology, www.aan.com; Oxford Centre for
systematic review of Maggin, Chafouleas, Goddard, and Johnson (2011) Evidence-Based Medicine, www.cebm.net).
2
into SCEDs to evaluate token economy interventions for students with Single-case methodology is defined as the intensive and prospective study
challenging behaviors found evidence of incomplete reporting. Of 24 of the individual in which (a) the intervention/s is manipulated in an experi-
mentally controlled manner across a series of discrete phases, and (b) mea-
eligible studies, a significant proportion failed to report on basic demo- surement of the behavior targeted by the intervention is made repeatedly (and,
graphic features of the participants, such as age (42%) and sex (33%). ideally, frequently) throughout all phases. Professional guidelines recommend
Moreover, 21% of the studies did not provide information on who that the experimental effect be demonstrated on at least three occasions by
implemented the intervention and 42% failed to specify the method used systematically manipulating the independent variable (Horner et al., 2005;
Kratochwill et al., 2010, 2013). This criterion helps control for the confound-
to record the data. Didden, Korzilius, van Oorsouw, and Sturmey (2006) ing effect of extraneous variables that may adversely affect internal validity
conducted a meta-analysis of 80 single-case reports on challenging be- (e.g., history, maturation), and allows a functional cause and effect relationship
havior in mild-to-moderate intellectual disability, finding that only 27% to be established between the independent and dependent variables.
 SCRIBE 2016 EXPLANATION AND ELABORATION 13

Figure 1. Common designs in the literature using a single participant. Reproduced from the expanded manual
for the Risk of Bias in N-of-1 Trials (RoBiNT) Scale (Tate et al., 2015) with permission of the authors; an earlier
version of the figure, taken from the original RoBiNT Scale manual (Tate et al., 2013a), was also published in
2013 (Tate et al., 2013b).

ioral interventions are often multicomponential and complex, and in that they clearly describe what was done. The suggested locations
some strategies to minimize bias, such as blinding, are difficult to in the article for the SCRIBE 2016 reporting items are not prescriptive
implement. For these reasons, two sets of reporting guidelines and authors should use their discretion about the most suitable loca-
were deemed necessary to cater to the different types of interven- tion.
tions and single-case methodologies used in the respective fields.
Similar reasoning drove the development of an extension of the
Methodology
CONSORT Statement for nonpharmacological RCTs (Boutron,
Moher, Altman, Schulz, & Ravaud, 2008). The CENT 2015 guideline is The first phase to develop the SCRIBE 2016 consisted of two
intended for use in medical N-of-1 trials, whereas the SCRIBE 2016 rounds of an online Delphi survey completed by SCED authors and
guideline is intended for SCEDs in the behavioral sciences. methodology experts, resulting in 44 items to be discussed at a
As with the Strengthening the Reporting of Observational Stud- consensus conference. At the meeting, held in Sydney, Australia, in
ies in Epidemiology (STROBE) Statement for observational stud- December 2011, participants reworked the items, resulting in a final
ies (Vandenbroucke et al., 2007), the SCRIBE 2016 is intended set of 26 items for the reporting guideline, which are described in this
to cover multiple research designs (specifically, the four most Explanation and Elaboration article. The SCRIBE 2016 Statement
common prototypical experimental designs: withdrawal/reversal, (Tate et al., 2016) provides a detailed description of the methodology
multiple-baseline, alternating/simultaneous treatments, and changing- of this process.
criterion designs, along with their variants, as well as adaptive designs).
Although we did not intend the SCRIBE 2016 to apply to non-SCEDs Results
using a single participant (designs shown below the solid horizontal line
in Figure 1), authors of such articles may find it useful to follow the This article provides examples of adequate reporting from the literature
guidance of those SCRIBE 2016 items as may apply to their study. for each of the 26 items, along with a rationale for inclusion of the item
Authors may also wish to consult the reporting guide for clinical CAse and, where available, evidence of bias resulting from incomplete report-
REports (CARE; Gagnier et al., 2014), the CONSORT for Social ing (the SCRIBE checklist of items appears in Table 1).
and Psychological Interventions (CONSORT-SPI; Montgomery et
al., 2013), and the Template for Intervention Description and Section 1: Title and Abstract (Items 1 and 2)
Replication (TIDieR) guideline (Hoffmann et al., 2014).
We prepared the SCRIBE 2016 within the CONSORT tradition of Item 1—Title: Identify the research as a single-case experimen-
guideline development (Moher, Schulz, Simera, & Altman, 2010b). tal design in the title.
Accordingly, the focus of this article concerns the reporting of stud-
Example.
ies, rather than education about the design of single-case experiments.
Graded exposure in vivo in the treatment of pain-related fear: a replicated
Moreover, in our use of examples from the literature to illustrate single-case experimental design in four patients with chronic low back
adequate reporting of each of the SCRIBE 2016 items, we adopt the pain. (Vlaeyen, de Jong, Geilen, Heuts, & van Breukelen, 2001, p. 151)
position of Boutron et al. (2008), wherein use of an example for a
specific item is not intended to imply that the study also provides Explanation. Although journals may place word limits on the
adequate examples of other items, nor even that the study per se is title, it is important to include as much information as possible in the
methodologically sound. Thus, although some of the examples may title, such as the intervention, the target behavior and the population.
not meet design standards, nonetheless they meet reporting standards In particular, the title should explicitly mention that the study is a
14 TATE ET AL.

Table 1
The Single-Case Reporting Guideline In BEhavioural Interventions (SCRIBE) 2016 Checklist

Item number Topic Item description

TITLE and ABSTRACT

1 Title Identify the research as a single-case experimental design in the title

2 Abstract Summarize the research question, population, design, methods including intervention/s (independent
variable/s) and target behavior/s and any other outcome/s (dependent variable/s), results, and
conclusions

INTRODUCTION

3 Scientific background Describe the scientific background to identify issue/s under analysis, current scientific knowledge, and
gaps in that knowledge base
4 Aims State the purpose/aims of the study, research question/s, and, if applicable, hypotheses

METHOD

DESIGN
5 Design Identify the design (e.g., withdrawal/reversal, multiple-baseline, alternating-treatments, changing-criterion,
some combination thereof, or adaptive design) and describe the phases and phase sequence (whether
determined a priori or data-driven) and, if applicable, criteria for phase change
6 Procedural changes Describe any procedural changes that occurred during the course of the investigation after the start of the
study
7 Replication Describe any planned replication
8 Randomization State whether randomization was used, and if so, describe the randomization method and the elements of
the study that were randomized
9 Blinding State whether blinding/masking was used, and if so, describe who was blinded/masked
PARTICIPANT/S or UNIT/S
10 Selection criteria State the inclusion and exclusion criteria, if applicable, and the method of recruitment
11 Participant characteristics For each participant, describe the demographic characteristics and clinical (or other) features relevant to
the research question, such that anonymity is ensured
CONTEXT
12 Setting Describe characteristics of the setting and location where the study was conducted
APPROVALS
13 Ethics State whether ethics approval was obtained and indicate if and how informed consent and/or assent were
obtained
MEASURES and MATERIALS
14 Measures Operationally define all target behaviors and outcome measures, describe reliability and validity, state how
they were selected, and how and when they were measured
15 Equipment Clearly describe any equipment and/or materials (e.g., technological aids, biofeedback, computer
programs, intervention manuals or other material resources) used to measure target behavior/s and other
outcome/s or deliver the interventions
INTERVENTIONS
16 Intervention Describe the intervention and control condition in each phase, including how and when they were actually
administered, with as much detail as possible to facilitate attempts at replication
17 Procedural fidelity Describe how procedural fidelity was evaluated in each phase
ANALYSIS
18 Analyses Describe and justify all methods used to analyze data

RESULTS

19 Sequence completed For each participant, report the sequence actually completed, including the number of trials for each
session for each case. For participant/s who did not complete, state when they stopped and the reasons
20 Outcomes and estimation For each participant, report results, including raw data, for each target behavior and other outcome/s
21 Adverse events State whether or not any adverse events occurred for any participant and the phase in which they occurred

DISCUSSION

22 Interpretation Summarize findings and interpret the results in the context of current evidence
23 Limitations Discuss limitations, addressing sources of potential bias and imprecision
24 Applicability Discuss applicability and implications of the study findings

DOCUMENTATION

25 Protocol If available, state where a study protocol can be accessed

26 Funding Identify source/s of funding and other support; describe the role of funders
 SCRIBE 2016 EXPLANATION AND ELABORATION 15

SCED because this differentiates the study from a case description. increased interest in verb retrieval and verb processing disturbances [ref].
The abstract is sometimes copyrighted by the journal, so the title may Unfortunately . . . One potentially useful speech pathology treatment for
be the only searchable information. Identifying the study as a SCED word production deficits involves the use of arm and hand gestures. . . .
in the title will ensure that the article is appropriately indexed for What remains to be developed is empirical evidence to support or refute
the suggestion that gesture is a potent treatment for verb retrieval deficits.
bibliographic databases (such as PsycINFO or Medline). Note that
This paper presents evidence . . . (Rose & Sussmilch, 2008, pp. 692– 693)
using “SCED” as a key word will not be sufficient for this purpose
because author key words are different from database key words and Explanation. The introduction is normally a discursive text that
may therefore not be searchable in electronic databases. overviews the relevant literature and identifies the gaps in knowledge
Item 2—Abstract: Summarize the research question, popula- that the current study aims to address. Ideally the text is succinct and
tion, design, methods including intervention/s (independent vari- targeted to the main issues that frame the context of the study. The
able) and target behavior/s and any other outcome measures introduction should commence with what is known about the problem
(dependent variable), results, and conclusions. area and its interventions, what is yet to be understood, and how this
study can address this gap.
Example.
This study tested the effectiveness of Imagery Rescripting (ImRs) for com- Item 4 —Aims: State the purpose/aims of the study, research
plicated war-related PTSD [posttraumatic stress disorder] in refugees. Ten question/s, and, if applicable, hypotheses.
adult patients in long-term supportive care with a primary diagnosis of
Example.
war-related PTSD and Posttraumatic Symptom Scale (PSS) score ⬎20 par-
The purpose of the present study was to examine the effects of a written
ticipated. A concurrent multiple baseline design was used with baseline
cueing treatment programme on verbal naming ability in two adults with
varying from 6 to 10 weeks, with weekly supportive sessions. After baseline,
aphasia. Treatment involved using a written cueing hierarchy, which was
a 5-week exploration phase followed with weekly sessions during which
modelled after CART [Copy and Recall Treatment] and included verbal and
traumas were explored, without trauma-focused treatment. Then 10 weekly
writing components. (Wright, Marshall, Wilson, & Page, 2008, p. 524)
ImRs sessions were given followed by 5-week follow-up without treatment.
Participants were randomly assigned to baseline length, and filled out the PSS
Explanation. The purpose and aims of the study need to be
and the BDI [Beck Depression Inventory] on a weekly basis. Data were
clearly described, normally at the conclusion of the Introduction.
analyzed with mixed regression. Results revealed significant linear trends
during ImRs (reductions of PSS and BDI scores), but not during the other These should take the form of research questions that define the
conditions. The scores during follow-up were stable and significantly lower independent variable, the dependent variable, and report if a formal
compared to baseline, with very high effect sizes (Cohen’s d ⫽ 2.87 (PSS) relationship was assessed. Statement of aims and, if applicable, hy-
and 1.29 [BDI]). One patient did clearly not respond positively, and revealed potheses, provide the reader with explicit directions regarding the way
that his actual problem was his sexual identity that he couldn’t accept. There in which the design, methods and results should be read, given that
were no dropouts. In conclusion, results indicate that ImRs is a highly these should all follow from the aims/research questions being asked.
acceptable and effective treatment for this difficult group of patients. (Arntz,
Sofi, & van Breukelen, 2013, p. 274)
Section 3: Method—Design (including both design structure, as
Explanation. The abstract needs to provide an accurate, infor- well as broader aspects of internal and external validity),
mative and unambiguous overview of the study. It is important that Participant/s, Context, Approvals, Measures and Materials,
all relevant information is included because many readers may not Interventions, Analysis (Items 5–18)
have access to the full article, or may choose to limit their reading
of the study to the abstract. The CONSORT guideline for abstracts Item 5—Design: Identify the design (e.g., withdrawal/reversal,
for randomized trials (Hopewell et al., 2008) provides useful multiple-baseline, alternating-treatments, changing-criterion,
information about how to write an abstract which, although written some combination thereof, or adaptive design) and describe the
for RCTs, has applicability to SCEDs. A structured abstract can be phases and phase sequence (whether determined a priori or data-
useful and make the abstract easy to follow. It is important that the driven) and, if applicable, criteria for phase change.
abstract clearly describes relevant features of the participant/s
Example 1: Withdrawal/reversal design.
(including clinical details where appropriate), defines the depen-
An A-B-A-B single-subject design evaluated a token economy for increasing
dent and independent variables, along with the SCED design used exercise in children with CF [cystic fibrosis] . . . Two advantages of this
to examine their relationship. The target behaviors and any addi- design are that it provides two evaluations of the treatment compared to
tional outcome measures (e.g., for generalization) used in the study baseline and it ends on a treatment phase, which is important from a clinical
should be specified, along with the way in which the target standpoint . . . The exercise diary data were used to determine when study
behavior is measured. An accurate summation of the outcomes of phase changes were made. The specific criteria were the following: (a) there
the study needs to be clearly detailed, along with disclosure of any were three or more stable data points, (b) there was a predictable pattern in
harms or adverse events, and conclude with a brief, cautious the data, or (c) there was no pattern, but the data points were predictably
appraisal of the significance of the research. random. (Bernard, Cohen, & Moffett, 2009, p. 354 –357)

Example 2: Multiple-baseline design.

Section 2: Introduction (Items 3 and 4) A randomized, concurrent, multiple-baseline single-case design was ap-
plied.13 Participants completed repeated measurements during a baseline
Item 3—Scientific background: Describe the scientific back- phase (phase A), an intervention period (phase B, 12 weeks) and a
ground to identify issue/s under analysis, current scientific knowl- postintervention period (phase A=). Phase A acted as a control and was
edge, and gaps in that knowledge base. therefore compared with phases B and A=. (Hoogeboom et al., 2012, p. 2)

Example (abbreviated). Example 3: Alternating-treatments design.

Verb production problems are an extremely common and pervasive The study used a multielement design with no baseline [ref] and a final
aphasic deficit following stroke. . . . Past research into word retrieval and “best treatment” phase [ref] to compare the effects of three contingent
production has mostly focused on nouns. . . . More recently there has been consequent events (Treatment A, adapted toys and devices; Treatment B,
16 TATE ET AL.

cause-and-effect commercial software; and Treatment C, instructor cre- vi. The number of trials within each session in a phase (i.e.,
ated video programs) on the frequency of stimulus activations. . . . occasions when the dependent variable is being measured)
Students received intervention in alternating treatments followed by the
best treatment phase, in which only the most effective intervention was vii. The duration of sessions
delivered. (Mechling, 2006, p. 98)
viii. The time interval between sessions (see also Item 20)
Example 4: Changing-criterion design.
[This study evaluated] a prompting and shaping intervention, in the form Additional descriptive information needs to be provided for spe-
of a changing criterion design . . . [to] assist this athlete in the technical cific design types.
development of his vault and corresponding height cleared . . . A photo-
electric beam . . . was set across the runway at a height of 2.30 m (5 cm Additional information required for multiple-baseline designs.
above the mean height obtained during baseline) . . . Intervention at this
height continued until the participant displayed stability in his perfor- i. The number of different (i.e., multiple) baselines (also referred
mance at a 90% level. Stability in performance referred to the successful to in the literature as data series, tiers, levels or legs) that the
repetition of three or more 90% performances. Following successful design contains. A graph alone is insufficient, because a graph
completion, intervention was also administered at the following heights: represents the results that were obtained, rather than the design as
2.35 . . . 2.52 m. (Scott, Scott, & Goldwater, 1997, pp. 573–574)
planned, which may change in response to the intervention (see Item
Explanation. In a broad sense, the design of a SCED encompasses 6).
all components of the methodology. For the purpose of this item, it
ii. Whether the baselines are across participants, behaviors or settings
specifically refers to the basic structure of phase sequencing and phase
onset used in the investigation. That is, the design defines how the iii. The method for determining treatment onset (e.g., response guided,
independent variable is manipulated, what the baseline conditions are, randomization), or, if necessary, that there was no specific rationale
and when/how the independent variable is introduced/changed. Manipu- or empirical basis
lation of these parameters allows the investigator to systematically control
the independent variable in order to demonstrate the experimental effect. iv. Sometimes multiple-baseline designs also incorporate either a
Moreover, using multiple phase changes in the design provides opportu- follow-up phase after the initial intervention phase or an “em-
nity for repeated demonstration of the experimental effect on the same bedded” design (e.g., alternating-treatments). When such a vari-
participant. This constitutes direct intrasubject replication (see Item 7 for ant is utilized, the complete sequence of phases should be clearly
intersubject replication and systematic replication). stated in the design description.
The design of SCEDs is crucial for determining the adequacy of
control of threats to internal validity and the experimental effect (Horner v. Whether or not the onset, and subsequent continuance, of data
et al., 2005; Kratochwill et al., 2013). Design characteristics need to be collection in each of the baselines occurred concurrently (i.e., at
clearly and specifically reported, otherwise it is difficult for the reader to the same points in time) or nonconcurrently. If nonconcurrent,
(a) determine if the study has sufficient experimental control to ade- provide a rationale for this choice.
quately establish a functional cause-effect relationship between the de-
pendent and independent variables and (b) evaluate the reliability of the Additional information required for alternating-treatments designs.
results. In a small random sample of 20 reports using a single participant
archived on the PsycBITE database, 45% of reports did not provide any i. Whether interventions were administered on the same day/
information on the type of design used and an additional 20% incorrectly session (e.g., Intervention 1 in the morning, Intervention 2 in
described the design (Tate et al., 2013b). the afternoon) or different days/sessions
Specific reporting requirements will vary among the experimental
ii. The way in which the order of the interventions was deter-
designs. A major strength of SCEDs is their flexibility and adaptability.
mined (e.g., randomized, counterbalanced, Latin square)
In addition to the “classic” designs described below, elements of these
designs can be creatively combined in a multitude of ways depending on iii. The detailed phase sequence (e.g., inclusion of a baseline
the research/clinical question being addressed and the intervention being preceding the intervention; a final “best treatment” phase fol-
considered (e.g., see Hayes, Barlow, & Nelson-Gray, 1999). A funda- lowing the intervention)
mental requirement in reporting SCEDs is that the basic structure of the
design is explicitly and accurately described. This consists of reporting Additional information required for changing-criterion designs.
the following eight invariant features that apply to all designs:
i. All criteria or decision rules used to determine when a phase
Basic information required for all designs, including withdrawal/
change occurs
reversal (A-B-A-B) design.
ii. Whether the criteria are set a priori or are response guided
i. The type of design (e.g., withdrawal/reversal)
Additional information required for adaptive designs. In these
ii. The number of phases (including baseline, experimental,
designs, the structure of the investigation (e.g., phase sequence/dura-
maintenance and follow-up phases)
tion, interventions, variations in intervention) is not fixed a priori, but
iii. The duration (length) of each phase depends, on an ongoing basis, on characteristics of the data (or
responses) from early (or preceding) phases, Authors should also
iv. The order in which the phases are sequenced (e.g., random- clearly describe the following:
ized, counterbalanced, data-driven)
i. Features or characteristics of the data (operationally defined)
v. The number of sessions in each phase that are used to regulate the study design
 SCRIBE 2016 EXPLANATION AND ELABORATION 17

ii. Which specific aspects of the study design, at each and all steps in Sidman, 1960). Direct intrasubject replication refers to replication of the
the study, are determined by which specific features of the data experimental effect within the design and addresses issues of internal
validity (see Item 5). The other two types of replication are relevant to this
Item 6 —Procedural changes: Describe any procedural changes item: (a) systematic replication (i.e., repeating the experiment with the
that occurred during the course of the investigation after the start same intervention but systematically changing characteristics of the in-
of the study. dividuals, setting, interventionists and/or behaviors) and (b) direct inter-
subject replication (i.e., repeating the same study but with additional
Example. individuals).
BST [behavioral skills training] was implemented with each child indi- Using a series of replications, Horner et al. (2005) and Kratochwill et
vidually in two 15- to 20-min sessions. . . . If the child did not obtain a al. (2010, 2013) note that it is possible to provide a strong basis for causal
score of 3 during an assessment after the initial BST sessions and two inference. They have proposed criteria for the purpose of establishing
booster sessions, an additional assessment session was turned into a evidence-based treatments: (a) a minimum of five methodologically
training session. . . . For Jake, who did not exhibit the correct behavior
strong research reports, (b) conducted by at least three different research
following BST or in situ training, an incentive phase was added. (Milten-
teams at three different geographical locations, and (c) with the combined
berger et al., 2004, pp. 514 –516)
number of cases being at least 20.
Explanation. There are occasions when, for a variety of rea- Authors should clearly and specifically state the number and type of
sons, the proposed implementation of a study changes from the replications in the Abstract and the Method section. If the study is
original plan. Changes may occur to any component of the study, using systematic replication to build an evidence-base, authors may
including (a) methodological design; (b) setting in which the study consider using the term systematic replication in the title. Authors
takes place; (c) target behavior and any other outcome measures, need to clearly indicate whether the replication refers to (a) the
with respect to content, method or frequency of administration; (d) replication of a previously published study or (b) intersubject repli-
equipment or materials to deliver the intervention; (e) use of cation within the current study.
practitioners and assessors; or (f) the intended intervention. Item 8 —Randomization: State whether randomization was
Changes involving participant attrition or early termination of used, and if so, describe the randomization method and the ele-
phases are addressed in Item 19. ments of the study that were randomized.
The researcher may actively initiate changes that are a departure from Example 1: Randomized sequence.
protocol or they may be thrust upon the researcher as a result of external Two treatments, one imitative and one cognitive-linguistic, were employed
factors. With respect to changes in the intervention, one of the reported and treatment order was determined randomly. (Leon et al., 2005, p. 96)
strengths of single-case methodology is the flexibility of implementation
of the intervention (Connell & Thompson, 1986; Gravetter & Forzano, Example 2: Randomized sequence (with restricted randomization).
2009). If adverse events occur or the intervention is not working suffi- A single case randomised experimental design with 12 phases was
ciently, then it is acceptable for the researcher to make alterations without used. Each phase lasted for one week. During six treatment phases,
necessarily compromising experimental control. participants wore the equipment and received cues (contingency elec-
trical stimulation). During six no treatment phases, participants wore
Authors need to report any changes or departure from the original
the equipment, but no cues were received. The phases were adminis-
plan or protocol, such as those listed above, along with reasons. They tered in a random order but always starting with a treatment phase, and
should also provide a statement about their impact on the interpreta- no more than two consecutive phases were the same. (Wenman et al.,
tion of the results. Places to describe procedural changes in the report 2003, p. 449)
will depend on the type of change/s, but either the Method or Dis-
cussion sections are appropriate. Example 3: Randomized onset.
The start of the treatment phase was determined randomly for each
Item 7—Replication: Describe any planned replication. participant, given the restriction that the baseline phase should last for at
least 6 weeks (42 days) and at most 12 weeks (84 days) . . . This means
Example 1: Systematic replication.
that the treatment phase could start on any day between the 42nd and the
We employed a multiple-baseline design to test the efficacy of a recently
84th days, resulting in a total of 43 possible assignments. (ter Kuile et al.,
developed approach for reducing school refusal behavior. . . . To maxi-
2009, p. 151)
mize external validity, the intervention was tested using a systematic
replication strategy, whereby only the major conceptual elements of the Explanation. The concept of randomization in SCEDs differs
intervention were retained from previous applications. . . . (Chorpita,
from its application in between-groups designs. In between-groups
Albano, Heimberg, & Barlow, 1996, p. 281)
designs, randomization exclusively refers to allocation of participants
Example 2: Direct intersubject replication. to intervention groups (i.e., experimental vs. control). By contrast, in
A single-subject ABAC design with replication across three participants SCEDs, it refers to (a) the random sequencing of baseline and inter-
was employed. (Leon et al., 2005, p. 96) vention phases, (b) the random determination of the commencement
time for each phase, and (c) the combined randomization of both
Explanation. Replication is a key feature of single-case method- phase order and phase starting point (Kratochwill & Levin, 2010). If
ology and is important because it has the capacity to inform whether more than one participant is being studied in withdrawal/reversal
and the extent to which an intervention is generalizable and hence is designs, individuals can also be randomly assigned to intervention
important for external validity. In spite of its central significance, conditions. In multiple-baseline designs, random allocation of partic-
replication is not commonly reported, with just over half of the studies ipants, behaviors or settings to each baseline of the design can also be
(54%) in the neurorehabilitation survey of Tate et al. (2014) describ- implemented. Ferron and Levin (2014) and Kratochwill and Levin
ing replication. In addition, although it may be evident that there is (2010) provide descriptions of these options.
replication in a study, it is not always explicitly stated. The sequencing of baseline and intervention phases in SCEDs
Three types of replication are described in the literature (Barlow et al., may be randomized using either simple (unrestricted) or blocked
2009; Gast, 2010a; Horner et al., 2005; Kratochwill et al., 2010, 2013; (restricted) randomization strategies (hence, randomized order de-
18 TATE ET AL.

signs). The commencement point in time of each phase may also be be used for blinding participants. By contrast, blinding/masking of
randomized using simple randomization (hence, randomized start assessors is usually feasible.
point designs). Randomization of both phase order and phase start In reporting SCEDs, if blinding was not implemented, authors
point can also be combined in any given design (Kratochwill & should state the reasons. When blinding was implemented, authors
Levin, 2010). Random allocation in SCEDs provides control of should clearly report who was blinded, and how the blinding was
potential confounders related to time (Edgington, 1996; Kratoch- achieved. Going beyond basic reporting standards, authors may wish
will & Levin, 2010; Onghena & Edgington, 2005) which addresses to consider reporting on procedures to assess the effectiveness of the
at least two potential sources of experimental bias in SCEDs: blinding procedures and the outcome.
history and maturation. Item 10 —Selection criteria: State the inclusion and exclusion
Researchers can also use randomization to assign specific stimulus criteria, if applicable, and the method of recruitment.
items to different stimulus sets (e.g., treated and nontreated stimuli),
although this type of randomization does not control for experimental Example.
bias related to history and maturation. We advertised the project in community newsletters and notices sent to
When randomization is used, authors should provide a reason for local hospitals and rehabilitation professionals who typically worked with
individuals with brain injuries. In the advertisements, we stated the
why it was used. They should also report specific details of (a) the
following inclusion criteria: parents must have a documented brain injury,
basic randomization strategy (i.e., simple or restricted) and (b) those
the children should be under the age of 10, and they must be demonstrat-
aspects of the design that were randomized (e.g., phase order, phase ing behavioral difficulties with the injured parent. Through subsequent
commencement, allocation of participants to interventions, allocation observation of the parent and child, we determined whether the child was
of participants, behaviors or settings to baselines). Authors need to demonstrating a serious level of oppositionality with the parent (e.g.,
describe any restrictions or modifications to the randomization pro- noncompliance to more than approximately 50% of parent-delivered
cess, which may be necessary for clinical or ethical reasons. requests). (Ducharme, Davidson, & Rushford, 2002, pp. 586 –587)
If randomization was not used, authors need provide a reason why
it was not used. They also need to report any decision criteria used to Explanation. Readers of SCEDs need to know as much as pos-
determine phase sequencing (e.g., counterbalancing), time points for sible about the participant(s), within the boundaries of anonymity,
phase onset (e.g., data driven), allocation of participants to interven- because, until generality has been demonstrated, results are only
tions (if applicable), and/or allocation of participants, behaviors or representative of the conditions under which the investigation was
settings to tiers. If decision criteria are based on participant-related conducted and for the individual/s who participated.
reasons (e.g., clinical considerations, severity of behavior, participant In situations where participants were actively recruited into the
needs), these need to be reported. study, inclusion and exclusion criteria should be provided. This in-
formation will assist with the identification of factors that may influ-
Item 9 —Blinding: State whether blinding/masking was used, ence a participant’s response to the intervention. It will also give an
and if so, describe who was blinded/masked. indication of the extent of the replicability and generalizability of
research findings (see also Item 11). The description of the selection
Example 1: Blind assessor.
process should provide detail regarding who was recruited, and also
The PQS [Psychotherapy Process Q-Sort] raters were not involved in
any other aspect of the study procedures, and had no prior information the way in which the participants were recruited (such as newspaper
regarding intended treatment approaches, design, or hypotheses. The advertisements, online recruitment targeting specific users, snowball-
raters were blind to the session number, treatment, and phase. Ratings ing methods via other study participants or relevant professionals,
of study tapes were made as part of PQS ratings of sessions from a distribution of brochures and leaflets).
larger sample [ref], and therefore were not rated consecutively or in Information should be provided about the way in which selection
comparison to one another. (Satir et al., 2011, p. 406) criteria were applied (e.g., by use of diagnostic instruments including
questionnaires and interviews). Details of methods, instruments ad-
No suitable examples of blinding of participant or practitioner in ministered, and classification or assessment criteria need to be clearly
the SCED behavioral sciences literature were identified. defined. Readers need to be able to unambiguously identify how
participant selection was accomplished for successful replication of
Explanation. Blinding (or masking) “refers to keeping trial par-
research (Horner et al., 2005).
ticipants, investigators (usually healthcare providers), or assessors
Method of recruitment may not be applicable in SCEDs in those
(those collecting outcome data) unaware of an assigned intervention,
situations where an individual presents with an issue that needs to
so that they are not influenced by that knowledge” (Schulz & Grimes,
be addressed. Such issues may be clinical (e.g., increasing a child’s
2002, p. 696). Lack of blinding or inadequate blinding in RCTs can
food intake) or nonclinical (e.g., improving an athlete’s perfor-
inflate estimates of intervention effects by up to 17%, especially if
mance). In these circumstances, authors should instead state the
trial outcomes involve subjective measures (Wood et al., 2008). It is
reasons that the individual presented to the service and the reason
not unreasonable to expect that this also applies to SCEDs. Blinding
for the intervention.
in SCEDs is reported very rarely, particularly for participants and
practitioners (Tate et al., 2013b). Item 11—Participant characteristics: For each participant, de-
Blinding is difficult to achieve in nonpharmacological trials involv- scribe the demographic characteristics and clinical (or other)
ing surgery, psychological interventions or rehabilitation (Bang, Ni, & features relevant to the research question, such that anonymity is
Davis, 2004; Boutron, Tubach, Giraudeau, & Ravaud, 2004). Given ensured.
that the majority of SCEDs involve nonpharmacological interven- Example.
tions, this is a pertinent issue. Although difficult, blinding of persons Three individuals with aphasia participated in the study. All had acquired
providing the intervention in nonpharmacological trials is not insur- aphasia secondary to a left hemisphere stroke. . . . See Table 1 for
mountable (e.g., Edinger, Wohlgemuth, Radtke, Marsh, & Quillian, participant’s demographic data. . . . Based on test performance and
2001). Boutron et al. (2007) provide a selection of strategies that can clinical judgment, all participants had nonfluent aphasia with good audi-
 SCRIBE 2016 EXPLANATION AND ELABORATION 19

tory comprehension. All participants exhibited significant word retrieval trolled manner. Multiple-baseline designs across settings are a case in
difficulties. Participants’ performances on the BNT [Boston Naming Test] point. Second, the detailed description of the relevant location and setting
were reviewed . . . P1’s word retrieval errors consisted of a mix of is central to the replicability of the study (see also Item 7). An indepen-
semantic (e.g., boat for canoe) and phonemic errors (e.g., fesmask for dent researcher may be interested in varying the location of the study as
mask) . . . (Rider, Wright, Marshall, & Page, 2008, pp. 162–163)
one step toward systematic replication. In this scenario, a sufficiently
Explanation. Inclusion of standard baseline participant character- detailed description of the setting is requisite information.
istics, including demographic information and functional status, en- Authors need to provide detailed information on the location of the
sures that the reader understands the presentation of the participants study, including the number and type of settings, as well the practitioners
and will be able to interpret the findings (Higginbotham & Bedrosian, or providers involved. There should be sufficient detail regarding the
1995). It is also important for generalization (Barlow et al., 2009) and location and setting of an intervention to enable others to evaluate how
facilitates meta-analysis of multiple studies (Robey, Shultz, Crawford, different this is from their own situation.
& Sinner, 1999). In spite of their relevance and importance, the Item 13—Ethics: State whether ethics approval was obtained
systematic review of Maggin, Chafouleas, et al. (2011) found that and indicate if and how informed consent and/or assent were
participant characteristics were incompletely reported, even for very obtained.
basic demographic features such as sex (not reported in 33% of
reports) and age (not reported in 42% of reports). Example.
The description of the participants of a study should include basic Approval for this research study was obtained from the human subjects
demographic information such as age, sex, ethnicity, socioeconomic research Internal Review Board at participating institutions. . . . If a
woman was interested, she was given a pamphlet with information about
status, geographic location, as well as diagnoses where indicated, and
the intervention and proposed research project before leaving the hospital,
functional or developmental abilities (Wolery & Ezell, 1993). Any or at a follow-up appointment. Women then contacted the primary inves-
diagnosis used should include instrumentation and scores. Lane, Wol- tigator (SMB) for more information and/or to schedule an initial assess-
ery, Reichow, and Rogers (2007) also recommend that baseline or ment appointment. An informed consent form was read, discussed and
environmental factors which serve to influence or maintain the par- signed before beginning the initial assessment and a copy of this consent
ticipant’s behavior during the initial baseline should be evaluated and form was given to the participant for their records. (Bennett, Ehrenreich-
reported. These features will go beyond simple description of sociode- May, Litz, Boisseau, & Barlow, 2012, p. 166)
mographic, medical and functional status variables.
Authors need to ensure that information provided does not lead to Explanation. It is virtually a universal requirement that research
identification of the participant. This is a particular risk in the treat- involving human participants requires review by and approval from an
ment of rare conditions. It is also important to protect an individual’s Institutional Ethics Committee. If the SCED was implemented as part
privacy if the study involves stigmatized conditions. of clinical care, ethics approval might not be required, as is the case
The participant in a SCED typically, but not always, represents an in N-of-1 trials (Punja, Eslick, Duan, Vohra, & the DEcIDE Methods
individual person. The participant, however, can also be a group whose Center N-of-1 Guidance Panel, 2014). Reporting whether ethics ap-
performance generates a single score per measurement period (such as proval has been obtained is not, however, a feature of all research
the rate of a particular behavior performed by all students within a reporting guidelines. For example, the CONSORT Statement (Moher
classroom during a set period of time). In that situation they are generally et al., 2010a) does not include this as a checklist item. Following the
referred to as a “case” or “unit.” It is important that the authors opera- CENT 2015 guidelines (Shamseer et al., 2015; Vohra et al., 2015), the
tionally define what constitutes the group and provide criteria for its SCRIBE 2016 also includes the reporting of ethical approval as a
selection (see Item 10). Authors need to specify the baseline character- checklist item for reporting SCEDs.
istics (to the same level of detail) for each participant (Wolery & Ezell, Written informed consent should always be secured (Mechling
1993) to allow the reader to ascertain the extent to which the results are & Gast, 2010). There may be instances when the participant cannot
generalizable (see also Item 24). provide informed consent (e.g., if the participant is a minor or
otherwise legally unable to provide informed consent). In this
Item 12—Setting: Describe characteristics of the setting and situation, their assent to participate should be sought, and consent
location where the study was conducted. also obtained from legal guardians or parents. It is not sufficient to
merely state informed consent/assent was obtained. Rather, the
Example.
The children and teacher comprised the full membership of a third grade
process by which consent/assent occurred needs to be described so
general education classroom in a rural postindustrial Northeast elemen- that it is clear who provided informed consent/assent. This is
tary school. . . . All observations were conducted in their homeroom class particularly important with vulnerable populations or where lim-
during the SSR [sustained silent reading] period . . . During SSR, students ited disclosure is necessary (National Health and Medical Research
sat at their desks, which were arranged in two rows of desks facing toward Council, 2009).
the teacher in the front of the room. (Methe & Hintze, 2003, p. 618)
Item 14 —Measures: Operationally define all target behav-
Explanation. It is critical to report the setting and location of the iors and outcome measures, describe reliability and validity,
study because these factors have implications for the generalizability state how they were selected, and how and when they were
and applicability of the findings (see Item 24). The context of the measured.
intervention will vary according to whether it is provided in primary,
Example 1: Operational definition of the target behaviors and
secondary or tertiary health care, classroom/educational facility or
how they were measured.
community settings. The location of the study will also vary according
Topographies of targeted behaviors for Bob included self-injury . . .
to whether the intervention is offered in urban, rural or remote Self-injury consisted of face slapping, defined as a forceful contact
locations, and this will have an impact on the service delivery model. between an open palm and cheek. . . . Bob also displayed spitting, defined
The setting is of particular interest for the reporting of SCEDs for two as spittle landing within 1 foot of another person. . . .
reasons. First, it may be an inherent a priori feature of the design to The primary dependent variable was the percent of intervals in which
introduce the independent variable across a range of settings in a con- maladaptive behaviors were observed during each 10-min session. Data
20 TATE ET AL.

were collected using paper and pencil during consecutive 10-s intervals validity and responsiveness of the instruments are reported. Any
cued by an audio tape throughout the entire session, resulting in a total of equipment used to measure the dependent variable/s should also have
60 consecutive intervals. Partial interval recording was used: during each established measurement properties, which, if available, should be
interval observers recorded whether or not any of the target maladaptive provided in the report.
behaviors had been observed for any portion of the interval.
When is it measured? A distinctive feature of SCEDs, in contrast
Observers underwent training in behavioral observation . . . and dem-
to group methodology, is that the target behavior (primary outcome
onstrated mastery prior to participating in the study. (Treadwell & Page,
1996, pp. 65– 66)
variable) is measured repeatedly and frequently throughout all phases,
including the baseline and intervention phases. Smith (2012, p. 519)
Example 2: Reliability of dependent variables when using non- observes that “the baseline measurement represents one of the most
standardized measures. crucial design elements of the SCED.” In spite of this, baseline data
Three different categories of interobserver agreement were calculated, as were not available for 22% of SCEDs in Smith’s systematic review.
follows. . . . Parent-therapist agreement: This category involved agreement Moreover, in other phases of the study, the number of data points
between compliance data coded by the parent and those coded live by the could not be readily identified.
research therapist. In this category, interobserver agreement was obtained on Target behavior/s need to be selected so that they are suitable for
39% of sessions conducted by parents, randomly selected from each of the repeated and frequent measurement. Previously, the recommended min-
phases across all children. Overall agreement averaged 92% for baseline imum number of data points per phase was three (Barlow & Hersen,
(range 82%–100%) and 98% for treatment, generalization, and follow-up
1984; Beeson & Robey, 2006), but more recently professional guidelines
sessions (range 94%–100%). (Ducharme et al., 2002, p. 588)
recommend a minimum of five data points per phase (Horner et al., 2005;
Explanation. A single item in the SCRIBE 2016 covers all as- Kratochwill et al., 2010, 2013). There should be a clear description of the
pects of the dependent variable/s: the what, how, and when of mea- number of sessions in which the target behavior is measured in each
suring the effect of the intervention. As with other research designs, phase, as well as the number of times it is measured in each session (i.e.,
the measurement process in SCEDs also needs to be valid and reliable. number of trials per session). Frequency of measurement of other out-
Validity is enhanced by selecting dependent variables that are (a) come variables depends on their role. Recommended practice is that
relevant to the behavior in question and that best match the interven- generalization measures are probed continuously throughout all phases
tion, as well as (b) accurate in their measurement, which is facilitated (Schlosser & Braun, 1994). By contrast, evaluation of social validity can
when the behavior that is targeted for intervention is operationally only logically occur after the intervention has taken place. As with target
defined. Reliability is enhanced when the behavior is measured in a behaviors, authors should clearly state the frequency and regularity of
manner that yields consistent results. measurement of all other outcome measures, including phases during
What is measured? SCEDs commonly use a variety of dependent which such measures were taken.
variables that play specific roles in the experiment. The primary outcome Item 15—Equipment: Clearly describe any equipment and/or
variable in SCED methodology is referred to as the target behavior. materials (e.g., technological aids, biofeedback, computer pro-
Target behaviors have three defining features in order to enhance quality grams, intervention manuals or other material resources) used to
of the study and minimize bias: they are specific, observable and repli- measure target behavior/s and other outcome/s or deliver the
cable (Barlow et al., 2009). Other dependent variables frequently used in interventions.
SCEDs may be considered akin to secondary outcome variables: Gener-
alization measures are increasingly recognized for their important role in Example 1: Software.
contributing to the external validity of the study (see also Item 24). In Training in use of email interface: “Participants 1– 4 used a mouse
addition, SCEDs have a strong tradition in promoting experiments that connected via USB port to activate the e-mail program. Participant 5
used a trackball instead of a mouse to accommodate his motor
address socially relevant behaviors and interventions. Additional mea-
impairment. . . . The instructor used a number pad . . . to control
sures are often incorporated into a SCED to specifically measure social presentation. . . . The program was run on . . . a laptop. . . . An altered
and ecological validity. interface was also developed to assess generalization to a slightly differ-
Authors need to provide operational definitions of the target be- ent platform. It included additional buttons . . . as well as rearrangement
havior, which should be objective, clear and complete (Kazdin, 2011), of existing buttons to novel positions.” (Ehlhardt, Sohlberg, Glang &
in order to convey what does and does not constitute an instance of the Albin, 2005, p. 571; p. 576)
dependent variable. In studies where there is more than one target
behavior, the report should clearly identify and describe each of the Example 2: Materials.
target behaviors in detail. Other dependent variables used in the study Photo Cue Cards [ref] were used as instructional stimuli. Four target
photographs and four control photographs were selected for each child.
(e.g., generalization measures, social validity measures) should also
Target stimuli are presented by child and instructional condition in Table
be described with equal clarity and precision.
II. A stopwatch was used to time the length of experimental sessions.
How is it measured? The “how” of measurement covers measure- (Holcombe, Wolery, & Snyder, 1994, p. 53)
ment procedures, including who selected and measured the target
behaviors and other outcome variables, along with their training in the Explanation. Target behaviors may be measured using behav-
assessment procedures. Authors also need to provide information on ioral observations and standardized scales (see Item 14) or, alterna-
the way in which the dependent variables were measured, justification tively, equipment. Similarly, many interventions used in the behav-
for the selection of those measures, and detail regarding what consti- ioral sciences are accompanied by materials and equipment. Complete
tutes a correct or incorrect response. and accurate reporting of equipment and materials is central to the
Because the target behaviors are highly specific to the presenting issues of replicability (see Item 7) and generalizability (see Item 24).
case in SCEDs, formal psychometric evaluation of the measures will A detailed description of the independent variable (i.e., the
generally not have been established. It is therefore recommended intervention) will include not only a description of the elements of
practice that evaluation of interobserver agreement on the target the intervention (see Item 16), but also any specific equipment
behavior is conducted and reported. When standardized instruments used, along with the way in which the equipment operates. Such
are used, it is essential that psychometric details regarding reliability, equipment will include training manuals, computer programs, bio-
 SCRIBE 2016 EXPLANATION AND ELABORATION 21

feedback techniques or any other materials required to implement observer watched videotapes of one training session from each step of the
the intervention. training programme (25% of sessions), and tallied the opportunities for the
When equipment is used to measure the dependent variable, the experimenter behaviours and the number of times the experimenter used
way in which it operates needs to be described, as well as its calibra- the expected behaviour (e.g., prompts, models, reinforcement). Adherence
to the protocol was calculated using the formula (EA ⫻ 100)/ET, where
tion. In addition, its measurement properties, if available, should be
EA ⫽ the experimenter behaviours. . . . Procedural reliability was 95% to
reported (see Item 14).
100% for each step. (Hickey, Bourgeois, & Olswang, 2004, p. 630)
Item 16 —Intervention: Describe the intervention and control
condition in each phase, including how and when they were Explanation. Wolery (1994) describes the collection of proce-
actually administered, with as much detail as possible to facilitate dural fidelity data as having three main functions: (a) to monitor the
attempts at replication. occurrence of relevant variables, (b) to provide documentation that
the experimental conditions occurred as planned, and (c) to provide
Example. information to practitioners about the use of the interventions. In spite
Check-in sessions were scheduled once a week during baseline to collect of the seriousness of unreliable implementation of the experimental
paperwork, monitor participant functioning, and to serve as an active waitlist conditions, fidelity checks in SCEDs in the behavioral sciences are
control condition. Following the baseline period, women entered the inter- infrequent (27% in the series of Didden et al., 2006). This result is
vention phase, which consisted of eight weekly sessions lasting approxi- comparable to the data on reporting quantitative measurement of the
mately 60 min. The therapist was the lead investigator of this study (SMB)
independent variable in three reviews of the contents of Journal of
who was a senior graduate student at the time of data collection. . . . Women
were told they could opt to include their partners in treatment sessions if they
Applied Behavior Analysis (Hagermoser Sanetti & Kratochwill, 2007–
wished . . . (Bennett et al., 2012, p. 166) 2008).
In recognition of the critical importance of the fidelity of the imple-
This description is accompanied by a detailed table (p. 165) de- mentation of study protocols, an item addressing adherence to the pro-
scribing the session, primary goal of each session, and brief descrip- tocol was introduced for the CONSORT Extension to Nonpharmacologi-
tion of session content and homework. cal Treatments (Boutron et al., 2008). Application of a prepared checklist
or steps of a protocol is the best way to document procedural fidelity (see
Explanation. Evidence of inadequate description of the intervention
Borrelli et al., 2005; Dane & Schneider, 1998; Gast, 2010b, pp. 99 –101).
abounds in the broader health-intervention research literature using group
Steps taken to evaluate procedural fidelity should be reported by authors,
methodology (e.g., Boutron et al., 2008; Dijkers et al., 2002; Glasziou,
including how it was measured and the results of its assessment. If
Meats, Heneghan, & Shepperd, 2008). The importance of describing the
procedural fidelity is evaluated during the course of the study, and found
independent variable itself in sufficient detail to allow replication has
to be suboptimal, authors may consider going beyond basic reporting
been emphasized in the general SCED literature.
standards to describe whether and what steps were taken to improve
The independent variable needs to be operationally defined, similar to
procedural fidelity (Hagermoser Sanetti & Kratochwill, 2014).
the descriptions of the target behavior/s and the outcome measures (see
Item 14). In situations where more than one intervention is used (e.g., Item 18 —Analysis: Describe and justify all methods used to
A-B-A-C-A-D, or alternating-treatments design) each intervention should analyze data.
be described. In addition to ancillary aspects of the intervention (i.e.,
Example 1: Use of visual analysis.
materials, manuals, stimulus items, equipment, software programs or
The split-middle technique [ref] was employed to detect changes in the
applications; see Item 15), the nature of the intervention per se needs to
number of successful shots within phases and resultant trend lines (Bar-
be clearly specified. This description includes information on who deliv- low & Hersen, 1984). White proposed that level, slope, and mean score of
ered the intervention and in which mode, whether it be individual, group, the celeration line (or trend) line be assessed as three descriptive analyses
distance, carer- or educator-focused, or whether telehealth or other tech- for conclusions. . . . Given that a point on the celeration line does not
nologies were used. It is critical to report the exact number, duration and actually explain the performance level, for brevity we have chosen to
frequency of the intervention sessions (Baker, 2012; Warren, Fey, & concentrate on the slope of the celeration line. (Mesagno, Marchant, &
Yoder, 2007). Specifically, intervention intensity or dosage should be Morris, 2009, p. 136)
described in terms of the following: (a) dose form (i.e., the typical task or
activity being used), (b) the dose (i.e., the number of times an active Example 2: Use of statistical analysis.
ingredient or teaching episode occurs per session), (c) the dose frequency As serial dependence . . . can bias the visual inspection,17 we checked
our data in each phase for serial dependence using the lag-1 method.12
(i.e., the number of intervention sessions per unit of time), (d) the total
If data were found to be significantly correlated, we transformed the
intervention duration (i.e., the total period of time in which an intervention is
data using a moving average transformation, in which the preceding
provided), and (e) the cumulative intervention intensity (i.e., the product of and succeeding measurements were taken into account.12,16 In addi-
dose by dose frequency by total intervention duration; Warren et al., 2007). tion, randomisation tests for multiple-baseline single-case designs
SCEDs typically compare the effect of an independent variable with were carried out. We expected phases B and A9 to be superior to phase
either a control condition or another independent variable (i.e., another A in terms of our health outcome assessment. Therefore, we tested the
intervention). The control condition, often called the baseline condition in null hypothesis that there would be no differential effect for any of
SCEDs, represents a period of time when the participant’s target behavior the measurement times using a randomisation test of the differences in
is recorded repeatedly before the intervention is introduced. The same the means between the preintervention phase and the intervention or
degree of specificity that is used to describe the experimental intervention postintervention phase.17 A p value of ⬍0.05 was considered statisti-
cally significant. For the premeasurements and postmeasurements, we
should also be used for the control condition.
considered change scores of 20% on validated questionnaires as clin-
Item 17—Procedural fidelity: Describe how procedural fidelity ically relevant.32 We used Stata/IC 10.1 for Windows for the descrip-
was evaluated in each phase. tive and visual analysis of the data and R version 2.14.1 for the
randomisation tests.31 (Hoogeboom et al., 2012)
Example.
Adherence to treatment procedures was accomplished by use of a manual that Explanation. Both visual and statistical techniques can be used
described the steps of the programme. . . . In addition, an independent, trained to analyze SCED data (see Appendix). They are considered com-
22 TATE ET AL.

plementary rather than mutually exclusive (Maggin & Odom, Explanation. A major benefit of SCEDs lies in their flexibility.
2014; Parker & Brossart, 2003; White, Rusch, Kazdin, & Hart- As described in Item 6, it is possible during the conduct of the
mann, 1989) and should, arguably, be used in combination (Davis investigation to change and fine-tune procedures and interventions
et al., 2013; Smith, 2012). that do not appear to be working. Any such deviation from the
Visual analysis relies upon visual inspection of the graphed data original plan of the study, however, needs to be clearly reported
to draw conclusions regarding the reliability and consistency of (see also Item 25). The present item pertains specifically to the
intervention effects (Lane & Gast, 2014). In the past, experts have report of procedural variations that reflect changes in any of phase
argued that visual analysis is the most sensitive and appropriate sequence, phase order, and/or number of sessions actually com-
way to detect intervention effects in SCEDs (e.g., Parsonson & pleted by each participant.
Baer, 1986). It has been the traditionally preferred and most Changes to phase sequence may occur in response to clinical or
frequently used approach (Busk & Marascuilo, 1992), but has ethical concerns, such as subsequently deciding to commence the
significant limitations (for discussion, see Lane & Gast, 2014; investigation with a treatment phase rather than the predetermined
Smith, 2012). Authorities have proposed guidelines for systematiz- randomized sequence. For similar reasons, a phase might be pre-
ing visual analysis (e.g., Kratochwill et al., 2010; Lane & Gast, maturely terminated, interrupted, extended, or substituted for rea-
2014), but there is not yet complete agreement about decision- sons of lack of efficacy, harms, periods of absence, and so forth.
making criteria to guide the process. Post hoc changes to the randomization schedule or the intended
Statistical analyses have advantages in that they (a) use an explicit set duration or structure of phases can significantly weaken the inter-
of operational rules and replicable methods, (b) provide a direct test of the nal validity of the study and therefore need to be reported.
null hypothesis, (c) utilize precisely defined criteria for significance, (d) Missing data due to attrition may bias the results, especially if
are useful when there is instability in the baseline or treatment effects are this reflects intentional or systematic noncompliance (Smith,
not well understood, and (e) can help to control for extraneous factors 2012). Moreover, if participants do not complete all phases as
(e.g., Kazdin, 1982a, 1982b). There is no “universal gold-standard,” planned, there may also be insufficient phase repetitions to dem-
however, and authors should choose the analytic method for SCEDs that onstrate adequate experimental control. Attrition of participants
is guided by a number of considerations: (a) design requirements and data within units (such as classrooms) over time may confound inter-
assumptions, (b) the research question being posed, (c) features of the vention effects especially if attrition is nonrandom and associated
data being analyzed, and (d) the interpretability, ease of computation and with the intervention itself.
proven validity of the technique for analysis of SCED data. For detailed In order to minimize such bias, Horner et al. (2005) argue that,
discussion of these issues, see Manolov, Gast, Perdices, and Evans (2014) regardless of attrition or missing data, results of any participant for
and Shadish (2014). If statistical analyses are used, authors need to report whom there is data for both a baseline and an intervention phase
whether underlying assumptions and other requirements pertinent to the should be reported. If techniques for dealing with missing data are
technique were evaluated for the data set being analyzed. This informa- used (e.g., retrospective data completion, such as completing dia-
tion will allow the reader to determine the suitability of the analytic ries), this also needs to be reported, given that such techniques can
methods used. introduce significant bias in the data due to incorrect recollection/
It is critically important that authors fully and clearly describe recall (Bolger, Davis, & Rafaeli, 2003). Accordingly, any changes
the method/s of analysis used, regardless of whether they select to the sequencing of phases or missing data should be reported,
visual, statistical or both techniques. Authors should state if the along with their reasons so that the reader can evaluate the integrity
method/s of analysis was prespecified before commencement of of the results and their interpretation.
data collection, and those changes (if any) that were subsequently
Item 20 —Outcomes and estimation: For each participant, re-
made as a consequence of limiting/problematic features of the data.
port results, including raw data, for each target behavior and
The rationale for selecting the analytic technique in terms of its
other outcome/s.
appropriateness to the study design and the research question
should be clearly stated, and the source reference describing the Example 1: Provide raw data (Figure 2).
technique should be cited. In the case of visual analysis, it is
important to clearly report the features of the data that were Example 2: Statistical analysis.
selected for analysis (along with reasons) and whether a systematic The three measures of treatment gains are shown in Table 2. Four
patients . . . demonstrated significant treatment gains as determined by
protocol was used and, if so, which one.
all three measures (C statistic, effect size, modified CDC [conservative
dual criteria]). One patient . . . did not demonstrate significant gains on
Section 4: Results—Sequence Completed, Outcomes and any measure. The patients who improved made variable gains in other
Estimation, Adverse Events (Items 19 –21) areas as indicated by a significant increase in their WAB [Western
Aphasia Battery] AQs [Aphasia Quotient] (mean increase ⫽ 5.58,
Item 19 —Sequence completed: For each participant, report the SD ⫽ 2.32, t ⫽ 4.81, df ⫽ 3, p ⬍ .05). (Crosson et al., 2009)
sequence actually completed, including the number of trials for
each session for each case. For participant/s who did not complete, Explanation. Traditionally, SCED results are reported in
state when they stopped and the reasons. graphical form which is, arguably, the clearest and most unambig-
uous way of depicting the major features of the raw data. At
Example: Deviation from protocol: Interruption of treatment. minimum, SCED data for each session should be reported in
Ms. O had 24 sessions of psychotherapy over a one hundred and 47 day graphic form. This does not preclude the additional reporting of
period, which included two periods of treatment interruption; once in the
raw data for each session (or for each trial within a session) for
middle of the BCT [Behavior Change Treatment] phase, and once at the end
of the BCT phase. Ms. O was randomly assigned to receive AFT [Alliance
each phase of the study in a tabulated format. Although tabular
Focused Treatment] for the first 4-week therapy phase, BCT for the second presentation is acceptable (and even desirable for verification in
4-week therapy phase, and AFT for the last 4-week therapy phase. The two meta-analysis), relevant features (e.g., consistency across similar
treatment interruptions coincided with the Thanksgiving and Christmas hol- phases) may not be directly obvious in this format. Alternatively,
idays, and occurred during BCT only. (Satir et al., 2011, p. 406) authors may provide information about where the raw numerical
 SCRIBE 2016 EXPLANATION AND ELABORATION 23

Figure 2. “Jason’s frequency of challenging behaviors, across settings.” Reprinted from “Structured Flexibil-
ity: The Use of Context-Sensitive Self-Regulatory Scripts to Support Young Persons With Acquired Brain Injury
and Behavioral Difficulties,” by T. J. Feeney, 2010, Journal of Head Trauma Rehabilitation, 25, p. 419.
Copyright 2015 by Wolters Kluwer Health. Reprinted with permission.

data set can be accessed. Raw data should be reported in the results other. More importantly, it also allows the reader to see how many
section for each measurement point/session in each phase of the sessions occurred in the initial A-phase of the second baseline (or
study for each participant, setting and target behavior/s. Even third, fourth, etc.) of the design after the intervention was intro-
though aggregation of data (e.g., averaging results over several duced in the first baseline.
sessions), may provide a clearer view of the apparent intervention Results of any statistical analyses for the target behavior/s and
effect, it may also mask or misrepresent various important features. other relevant outcome variables also need to be reported. This
Such features may include (a) stability of the initial baseline phase, report should be done for each participant and setting in the study.
(b) variability and trends within a phase, (c) degree of consistency Irrespective of the analysis used, authors need to clearly state
between similar phases (e.g., intervention phases), (d) the degree which phases and features of the data were compared. Any changes
of overlap between baseline and intervention phases, (e) magnitude or deviations in a preplanned analysis strategy should also be
of effect latency following intervention phase onset. Readers need reported, as well as the reasons that it was necessary to make such
to be able to critically evaluate all these aspects of the data. They
changes. For statistical analyses, the value of the calculated sta-
also need to be able to draw their own conclusions about how
tistic/s, standard errors, and any associated probability level should
adequately the investigators have taken any anomalies into account
also be reported.
when appraising the clinical value of the intervention (Barlow et
al., 2009). Item 21—Adverse events: State whether or not any adverse
The metric used on the horizontal axis of graphed data should be events occurred for any participant and the phase in which they
in units of real time (i.e., days, weeks, etc.) rather than session occurred. The current reporting of adverse events is rare in SCEDs
number. This information allows the reader to more clearly inter- where behavioral interventions have been applied. Only a single
pret and critically evaluate the results of the study. Providing an example (Hoogeboom et al., 2012) was identified where authors
exact chronology of the time interval between sessions allows the mentioned that adverse events were monitored, but no information
reader to accurately evaluate patterns of consistency between sim- was provided on the way in which adverse effects were measured, nor
ilar phases and effect latency following intervention onset. Carr were results provided.
(2005) argues that using a real-time metric on the horizontal axis
is particularly relevant in multiple-baseline designs. The reason is Explanation. The reporting of adverse events or harms is rare
so that the reader can determine the order in which sessions across in SCEDs of behavioral interventions. Their report is a more
participants, settings or behaviors were conducted relative to each common feature in medical interventions in randomized and non-
24 TATE ET AL.

randomized trials and observational studies (Golder, Loke, & larly pertinent to SCEDs in instances where findings are unclear or
Bland, 2011), even though it is often inadequate (Papanikolaou, adequate replication has not occurred (see Item 7).
Christidi, & Ioannidis, 2006; Vandenbroucke, 2006). This problem In terms of interpreting the study findings in the context of the
is despite the reporting of harms being a specific checklist item in current evidence from the literature, the CONSORT Statement
the CONSORT 2010 Statement (Schulz, Altman, & Moher, 2010). (Schulz et al., 2010) suggests that results of clinical trials are inter-
Although there are no clear definitions of harms, the CONSORT preted with respect to the knowledge base, as synthesized in system-
Extension for Better Reporting of Harms in RCTs (Ioannidis et al., atic reviews. In some research areas using SCEDs, meta-analyses are
2004) has recommended harms as the preferred term to describe an available and it is recommended that information from these be
adverse event (as opposed to describing the “safety” of a treat- incorporated when available (as in the above example of Lo et al.,
ment). Particular recommendations include that (a) authors make 2011, referring to the meta-analysis of Chard, Ketterlin-Geller, Baker,
specific mention of harm in the title or abstract, as well as the Doabler, & Apichatabutra, 2009).
introduction where harms are a primary outcome measure; and (b)
Item 23—Limitations: Discuss limitations, addressing sources
there is specification of the harms in reporting of results, with
of potential bias and imprecision.
special attention paid to discontinuations and withdrawals due to
adverse events. The guideline also recommends that authors should Example.
(a) present the absolute risk of each adverse event (specifying type, The validity of comparisons between AFT [Alliance Focused Treatment]
grade, and seriousness per arm of the trial), (b) describe any and BCT [Behavioral Change Treatment] was compromised by several
subgroup analysis and exploratory analysis for harms, and (c) factors, including the significant intervention interruptions during BCT,
provide a balanced discussion of benefits and harms with emphasis the administration of psychotropic medication during the study. . . .
on study limitations, generalizability and other sources of infor- Furthermore, there may have been significant carry-over effects from one
phase to another, as skills learned during one phase could not be “un-
mation on harms (Ioannidis et al., 2004). The guide can be adapted
learned” . . . Limitations to measurement include possible limitation to the
to accommodate SCED methodology. accuracy of the self-report nature of the kilocalorie intake, which was not
The SCRIBE 2016 recommends that authors make full and explicit verified by other sources of data collection (e.g., independent observa-
disclosure of any harms or adverse events that occurred to any tion). (Satir et al., 2011, p. 417)
participant during the course of the SCED trial, including the absence
of these events. Loke and colleagues (2007) suggest a framework to Explanation. It is often tempting for authors to focus on the
enable a systematic, manageable and clinically useful way to define positive findings of their results and to glide over the flaws in their
adverse effects. It includes a predefined classification of adverse study. However, all studies have limitations that can either bias or
effects as diagnosed by the clinician, by test results, or by participant- confound results. Important limitations reflect threats to the validity of
reported symptoms (e.g., pain). the study that introduce potential bias in the findings.
There are many potential limitations in SCED studies that compro-
mise the extent to which results are unbiased, reliable and likely to
Section 5: Discussion—Interpretation, Limitations, Applica- generalize (e.g., see Horner et al., 2005; Tate et al., 2013b; Wolery,
bility (Items 22–24) Dunlap & Ledford, 2011). These limitations include the following: (a)
poor matching of the design to the type of intervention (see Item 5);
Item 22—Interpretation: Summarize findings and interpret the
(b) inadequate replication (see Items 7 and 24); (c) absence of blind-
results in the context of current evidence.
ing (see Item 9); (d) lack of randomization (see Item 8); (e) impreci-
Example. sion with respect to the description of the participant’s functional
Results showed that the repeated reading program combining several abilities (and, where applicable, diagnosis), making it difficult for
research-based components . . . improved fluency on second-grade trans- readers to generalize to other cases (see Items 10 and 11); (f) problems
fer passages for the three participants lending support to the existing with the operational definition of the target behavior or assessment of
literature on repeated reading [ref]. . . . With research on repeated reading its reliability, insufficient number of data points in some or all of the
spanning decades and numerous studies demonstrating successful out- phases to meet minimum standards (see Item 14); (g) absence of
comes . . . this practice holds great promise as a strategy for improving
information regarding procedural fidelity (see Item 17); and (h) reli-
reading fluency. However, as suggested by [the metaanalysis of] Chard
and colleagues (2009), the current research literature on repeated reading
ance on visual analysis for ambiguous cases, insufficient number of
is not sufficient for it to be designated as an evidence-based practice. (Lo, data points required for specific statistical analyses (e.g., the C sta-
Cooke, & Starling, 2011, pp. 133, 136) tistic requires at least eight data points per phase), use of statistical
procedures that do not deal adequately with extant features of the data
Explanation. An early section of the discussion needs to provide (e.g., trend, variability or auto-correlation), or whose underlying as-
a clear and concise summary of the findings of the study, including the sumptions are not met (see Item 18).
strength of the intervention effect and the clinical importance of the In terms of adequate reporting, authors need to provide a systematic
findings. The results should be interpreted in terms that are specific to discussion of specific limitations associated with their findings (as de-
the study, as well as more generally with reference to the current scribed above) which is also contextualized within the relevant literature.
literature. Interpretation specific to the study will benefit from taking In this way the reader is provided with a realistic, critical appraisal of the
into account the aims of the study, along with the robustness of contribution that the study makes to the field and the extent to which the
methodology and procedures. Item 23 on limitations of the study is a results reflect a strong finding that is likely to be repeated.
separate item in the SCRIBE 2016 Statement, but has obvious rele-
Item 24 —Applicability: Discuss applicability and implications
vance to the item on interpretation.
of the study findings.
Vandenbroucke and colleagues (2007) note that overinterpretation
is a common problem in observational studies. For the STROBE Example.
Statement, they advocate that caution is exercised in interpreting the The results of the present study replicate the findings of [ref] with regard
findings of a study. A cautious approach to interpretation is particu- to the effect of using DRA [differential reinforcement of alternative
 SCRIBE 2016 EXPLANATION AND ELABORATION 25

behavior], nonremoval of the fork, and stimulus fading to increase variety also included in the CENT 2015 Statement (Shamseer et al., 2015;
of food intake. The study extends previous findings by showing that the Vohra et al., 2015). Moher et al. (2010a, p. 21) indicate that the
intervention package was effective independent of who fed the child. protocol refers to the planned methods of the “complete trial (rather
The effect of our treatment on John’s consumption of nonpreferred than a protocol of a specific procedure within a trial).” The rationale
foods did not generalize across settings in the absence of intervention
for including the protocol item in the CONSORT 2010 Statement was
in the home, but multisetting training led to transfer across settings and
caregivers. Our study also shows that the treatment package described
based on published evidence of the discrepancy between the trial
by [ref] was effective during typically occurring mealtimes with protocol and the subsequent published report (e.g., selective reporting
regularly scheduled food types, and that the treatment was effective for of results, post hoc change in the main outcome measure). Such
increasing the number and variety of originally nonpreferred foods. discrepancies continue to be documented (Dwan et al., 2011).
(Valdimarsdottir, Halldorsdottir, & SigurÐardóttir, 2010, p. 105) Having a protocol readily accessible makes authors accountable for
any changes that are made to the planned research design and analysis.
Explanation. The concept of applicability or generality is As noted, however, a special feature of SCED methodology is its
based on the assumption that inferences can be drawn from the flexibility in terms of modifying the design and/or intervention after
condition in which an intervention effect was demonstrated, to the trial has commenced (e.g., if the participant does not respond to
other conditions based on known similarities and differences be- intervention or specific research problems arise). These modifications
tween these conditions (Gast, 2010a).
do not necessarily compromise the experiment (Connell & Thompson,
The reader should be provided with a discussion of implications of
1986; Gravetter & Forzano, 2009). As a consequence, in a SCED the
(a) conceptual/theoretical considerations, (b) clinical/practical consid-
design and/or intervention actually received may depart from an a
erations, and (c) methodological considerations, which impact on the
priori protocol. Such departure is considered acceptable within SCED
generalizability of the findings. Conclusions could be drawn, for
methodology, as long as the authors declare such departure and
example, from information provided about replication (Item 7), inclu-
provide justification for it (see Item 6). Nonetheless, the flexibility to
sion and exclusion criteria (Item 10), participant characteristics (Item
modify the design and intervention in SCEDs does not imply that an
11), and generalization measures (Item 14).
a priori protocol is not relevant.
The replication process (see also Item 7) involves an increasing
The SPIRIT 2013 Statement (Standard Protocol Items: Recommen-
number of variations to the different dimensions (most importantly
dations for Interventional Trials; Chan et al., 2013) provides guidance
participants, setting and practitioner) that can be changed with every
for the report of protocols of clinical trials and can be used as a guide
replication (Barlow et al., 2009; Sidman, 1960). Each replication will
for preparing a protocol on a SCED study. As noted in the CONSORT
add information regarding the generalizability of the findings. The
stage of the replication process (direct or systematic; see, e.g., Gast, 2010 Explanation and Elaboration document (Moher et al., 2010a),
2010a) should be made clear. Similarities and differences between the there are many ways in which a protocol can be made available,
current and previous studies need to be explicated. including trial registries, journals that publish protocols, website of
Reference should be made to factors that determined baseline the journal publishing the main results of the study, author’s institu-
performance, such as described in Items 10 and 11, because these tional website, contact with the author. Reference to the study proto-
factors are hypothesized to influence relations between the indepen- col can be made either in the text or as a footnote.
dent variable and the dependent variable in a very specific way Item 26 —Funding: Identify source/s of funding and other sup-
(Horner et al., 2005; Wolery, et al. 2011). If outcome measures port; describe the role of funders.
additional to the target behavior (Item 14) are used for the purpose of
generalization, authors should discuss the evidence to support (or not) a Example.
functional relationship between the independent variable and the gener- Funding.
alization variable in the context of a theoretical framework, if available. The study was financed by the Sint Maartenskliniek Nijmegen and
If responses to the intervention differed across participants (or between Woerden, the Netherlands.
studies), reasons and proposed causal relationships of this finding should
be discussed and new propositions made to explain the finding. Competing interests.

All authors declare: no support from any organisation for the submitted work,
Section 6: Documentation—Protocol, Funding (Items 25–26) no financial relationships with any organisations that might have an interest
Item 25—Protocol: If available, state where a study protocol in the submitted work in the previous 3 years and no other relationships or
can be accessed. No published SCED studies were identified that activities that could appear to have influenced the submitted work. (Hooge-
boom et al., 2012, p. 8)
contained information on how to access a protocol. A number of
published reports indicated that the research protocol of a SCED was Explanation. Journals often request that authors disclose funding
reviewed by an institutional research committee, but this pertains to sources. It is important that readers of the manuscript can make a
Item 13. Several SCED protocols were identified in trial registries judgment as to whether the funders have control over knowledge
(Lloyd at www.anzctr.org.au, Trial ID ACTRN12611000812998; dissemination or whether they provided funds but the researchers
Pool at www.anzctr.org.au, Trial ID ACTRN12611000531910; Wam-
worked independently and autonomously. A consistently and over-
baugh, Mauszycki, Cameron, Wright, & Nessler, 2013; Wambaugh,
whelmingly strong association between industry support and proin-
Nessler, C& Wright, 2013, at www.clinicaltrials.gov) but personal
dustry findings in group clinical trials has been demonstrated (Sis-
communication with the authors indicated that the studies were not yet
mondo, 2008a, 2008b). The likelihood of finding a positive result
published (Lloyd & Sherrington, 2011) or the published report did not
when backed by industry funding has been estimated in the range of
make reference to the protocol (Pool, Blackmore, Bear, & Valentine,
3.6 to 4.05 times greater than when not (Bekelman, Li, & Gross, 2003;
2014; Wambaugh, Mauszycki, et al., 2013; Wambaugh, Nessler, et al.,
Lexchin, Bero, Djulbegovic, & Clark, 2003). Although methodolog-
2013).
ical quality is not necessarily compromised in industry-funded re-
Explanation. Trial protocol availability was a new item intro- search (Sismondo, 2008a), bias may infiltrate in other ways such as
duced for the CONSORT 2010 Statement (Schulz et al., 2010) and is the decision to use less active controls (Bekelman et al., 2003).
26 TATE ET AL.

Provision of a grant number can be helpful because it enables readers trials. The CONSORT Statement. Journal of the American Medical
to retrieve the details of the grant. Otherwise, one needs to know the Association, 276, 637– 639. http://dx.doi.org/10.1001/jama.1996
following: Did the funder provide equipment or money? Did the .03540080059030
funder have control over where the study was published or what was Bekelman, J. E., Li, Y., & Gross, C. P. (2003). Scope and impact of financial
published? Were the funders the authors, or did they edit the manu- conflicts of interest in biomedical research: A systematic review. Journal of
the American Medical Association, 289, 454 – 465. http://dx.doi.org/10
script? Any conflicts of interest should be stated explicitly. This
.1001/jama.289.4.454
information can be provided in the body of the text or in a footnote.
Bennett, S. M., Ehrenreich-May, J., Litz, B. T., Boisseau, C. L., & Barlow,
D. H. (2012). Development and preliminary evaluation of a cognitive-
Conclusion behavioral intervention for perinatal grief. Cognitive and Behavioral Prac-
We developed the SCRIBE 2016 to assist investigators in the tice, 19, 161–173. http://dx.doi.org/10.1016/j.cbpra.2011.01.002
Beretvas, S. N., & Chung, H. (2008). A review of meta-analyses of single-
behavioral sciences to report SCEDs with transparency, accuracy,
subject experimental designs: Methodological issues and practice. Evidence-
clarity and completeness. This article provides rationale for and ex-
Based Communication Assessment and Intervention, 2, 129 –141. http://dx
planation of the 26 SCRIBE 2016 items. It also includes examples of .doi.org/10.1080/17489530802446302
adequate reporting of specific SCRIBE 2016 items, drawing on arti- Bernard, R. S., Cohen, L. L., & Moffett, K. (2009). A token economy for
cles in the published literature. Authors reporting on one specific type exercise adherence in pediatric cystic fibrosis: A single-subject analysis.
of single-case methodology, the medical N-of-1 trial with multiple Journal of Pediatric Psychology, 34, 354 –365. http://dx.doi.org/10.1093/
cross-overs, will find it helpful to consult the CENT 2015 Statement jpepsy/jsn101
(Shamseer et al., 2015; Vohra et al., 2015), which was developed for Bloom, M., & Fischer, J. (1982). Evaluating practice: Guidelines for the
that particular methodology. We welcome feedback from users of the accountable professional (1st ed.). Englewood Cliffs, NJ: Prentice Hall.
SCRIBE 2016, which can be made through the SCRIBE website Bolger, N., Davis, A., & Rafaeli, E. (2003). Diary methods: Capturing life as
(www.sydney.edu.au/medicine/research/scribe). it is lived. Annual Review of Psychology, 54, 579 – 616. http://dx.doi.org/10
Since the first CONSORT guideline appeared in 1996 (Begg et al., .1146/annurev.psych.54.101601.145030
Borrelli, B., Sepinwall, D., Ernst, D., Bellg, A. J., Czajkowski, S., Breger, R.,
1996), medical journals have continued advocating the use of pre-
. . . Orwig, D. (2005). A new tool to assess treatment fidelity and evaluation
scriptive reporting guidelines in the CONSORT tradition. The
of treatment fidelity across 10 years of health behavior research. Journal of
EQUATOR network (www.equator-network.org) is a useful resource Consulting and Clinical Psychology, 73, 852– 860. http://dx.doi.org/10
to keep up-to-date with new developments in this field. This develop- .1037/0022-006X.73.5.852
ment has not occurred in the behavioral sciences to the same degree, Boutron, I., Guittet, L., Estellat, C., Moher, D., Hróbjartsson, A., & Ravaud, P.
although many authors of intervention studies in the behavioral sciences (2007). Reporting methods of blinding in randomized trials assessing non-
consult relevant CONSORT Statements (e.g., CONSORT Extension to pharmacological treatments. PLoS Medicine, 4, e61. http://dx.doi.org/10
Nonpharmacological Interventions; Boutron et al., 2008). The influence .1371/journal.pmed.0040061
of CONSORT is such that the peak medical journals require that Boutron, I., Moher, D., Altman, D. G., Schulz, K. F., & Ravaud, P., & the
authors address all of the criteria of the relevant guideline in their CONSORT Group. (2008). Extending the CONSORT Statement to random-
report. The benefit has resulted in improved reporting (Turner et al., ized trials of nonpharmacologic treatment: Explanation and elaboration.
2012). It will be advantageous to the behavioral sciences if journals Annals of Internal Medicine, 148, 295–309. http://dx.doi.org/10.7326/0003-
4819-148-4-200802190-00008
publishing single-case methodology also endorse use of the SCRIBE
Boutron, I., Tubach, F., Giraudeau, B., & Ravaud, P. (2004). Blinding was
2016 in this way. judged more difficult to achieve and maintain in nonpharmacologic than
pharmacologic trials. Journal of Clinical Epidemiology, 57, 543–550. http://
References dx.doi.org/10.1016/j.jclinepi.2003.12.010
Allison, D. B., & Gorman, B. S. (1993). Calculating effect sizes for meta- Box, G. E., & Jenkins, G. M. (1970). Time series analysis: Forecasting and
analysis: The case of the single case. Behaviour Research and Therapy, 31, control (1st ed.). San Francisco, CA: Holden-Day.
621– 631. http://dx.doi.org/10.1016/0005-7967(93)90115-B Bulté, I., & Onghena, P. (2012). When the truth hits you between the eyes: A
Alresheed, F., Hott, B. L., & Bano, C. (2013). Single subject research: A software tool for the visual analysis of single-case experimental data. Meth-
synthesis of analytic methods. Journal of Special Education Apprenticeship, odology: European Journal of Research Methods for the Behavioral and
2, 1–18. http://josea.info/archives/vol2no1/vol2no1-1-FT.pdf Social Sciences, 8, 104 –114. http://dx.doi.org/10.1027/1614-2241/a000042
Arntz, A., Sofi, D., & van Breukelen, G. (2013). Imagery Rescripting as Busk, P., & Marascuilo, L. A. (1992). Statistical analysis in single-case
treatment for complicated PTSD in refugees: A multiple baseline case series research: Issues, procedures, and applications to multiple behaviors. In T. R.
study. Behaviour Research and Therapy, 51, 274 –283. http://dx.doi.org/10 Kratochwill & J. R. Levin (Eds.), Single-case research design and analysis:
.1016/j.brat.2013.02.009 New directions for psychology and education (pp. 159 –186). Hillsdale, NJ:
Baker, E. (2012). Optimal intervention intensity. International Journal of Erlbaum.
Speech-Language Pathology, 14, 401– 409. http://dx.doi.org/10.3109/ Carr, J. E. (2005). Recommendations for reporting multiple-baseline designs
17549507.2012.700323 across participants. Behavioral Interventions, 20, 219 –224. http://dx.doi
Bang, H., Ni, L., & Davis, C. E. (2004). Assessment of blinding in clinical .org/10.1002/bin.191
trials. Controlled Clinical Trials, 25, 143–156. http://dx.doi.org/10.1016/j Center, B. A., Skiba, R. J., & Casey, A. (1985). A methodology for the
.cct.2003.10.016 quantitative synthesis of intra-subject design research. The Journal
Barlow, D. H., & Hersen, M. (1984). Single case experimental designs. Strate- of Special Education, 19, 387– 400. http://dx.doi.org/10.1177/
gies for studying behaviour change (2nd ed.). Boston, MA: Allyn & Bacon. 002246698501900404
Barlow, D. H., Nock, M. K., & Hersen, M. (2009). Single case experimental Chan, A-W., Tetzlaff, J. M., Altman, D. G., Laupacis, A., Gøtzsche, P. C.,
designs: Strategies for studying behavior change (3rd ed.). Boston, MA: Krleža-Jerić, K., . . . Moher, D. (2013). SPIRIT 2013 Statement: Defining
Pearson. standard protocol items for clinical trials. Annals of Internal Medicine, 158,
Beeson, P. M., & Robey, R. R. (2006). Evaluating single-subject treatment 200 –207. http://dx.doi.org/10.7326/0003-4819-158-3-201302050-00583
research: Lessons learned from the aphasia literature. Neuropsychology Chard, D. J., Ketterlin-Geller, L. R., Baker, S. K., Doabler, C., & Apichatabu-
Review, 16, 161–169. http://dx.doi.org/10.1007/s11065-006-9013-7 tra, C. (2009). Repeated reading intervention for students with learning
Begg, C., Cho, M., Eastwood, S., Horton, R., Moher, D., Olkin, I., . . . Stroup, disabilities: Status of the evidence. Exceptional Children, 75, 263–281.
D. F. (1996). Improving the quality of reporting of randomized controlled http://ecx.sagepub.com/content/75/3/263.short
 SCRIBE 2016 EXPLANATION AND ELABORATION 27

Chorpita, B. F., Albano, A. M., Heimberg, R. G., & Barlow, D. H. (1996). A Ferron, J., & Onghena, P. (1996). The power of randomization tests for
systematic replication of the prescriptive treatment of school refusal behav- single-case phase designs. Journal of Experimental Education, 64, 231–239.
ior in a single subject. Journal of Behavior Therapy and Experimental http://dx.doi.org/10.1080/00220973.1996.9943805
Psychiatry, 27, 281–290. http://dx.doi.org/10.1016/S0005-7916(96)00023-7 Ferron, J., & Ware, W. (1994). Using randomization tests with responsive
Connell, P. J., & Thompson, C. K. (1986). Flexibility of single-subject exper- single-case designs. Behaviour Research and Therapy, 32, 787–791. http://
imental designs. Part III: Using flexibility to design or modify experiments. dx.doi.org/10.1016/0005-7967(94)90037-X
The Journal of Speech and Hearing Disorders, 51, 214 –225. http://dx.doi Fisher, R. A. (1920). A mathematical examination of the methods of deter-
.org/10.1044/jshd.5103.214 mining the accuracy of an observation by the mean error, and by the mean
Crosson, B., Moore, A. B., McGregor, K. M., Chang, Y.-L., Benjamin, M., square error. Monthly Notices of the Royal Astronomical Society, 80, 758 –
Gopinath, K., . . . White, K. D. (2009). Regional changes in word-production 770. http://dx.doi.org/10.1093/mnras/80.8.758
laterality after a naming treatment designed to produce a rightward shift in Friedman, M. (1937). The use of ranks to avoid the assumption of normality
frontal activity. Brain and Language, 111, 73– 85. http://dx.doi.org/10.1016/ implicit in the analysis of variance. Journal of the American Statistical
j.bandl.2009.08.001 Association, 32, 675–701. http://dx.doi.org/10.1080/01621459.1937
Dane, A. V., & Schneider, B. H. (1998). Program integrity in primary and early .10503522
secondary prevention: Are implementation effects out of control? Clinical Gabler, N. B., Duan, N., Vohra, S., & Kravitz, R. L. (2011). N-of-1 trials in the
Psychology Review, 18, 23– 45. http://dx.doi.org/10.1016/S0272- medical literature: A systematic review. Medical Care, 49, 761–768. http://
7358(97)00043-3 dx.doi.org/10.1097/MLR.0b013e318215d90d
Davis, D. H., Gagné, P., Fredrick, L. D., Alberto, P. A., Waugh, R. E., & Gagnier, J. J., Kienle, G., Altman, D. G., Moher, D., Sox, H., & Riley, D., &
Haardörfer, R. (2013). Augmenting visual analysis in single-case research the CARE Group. (2014). The CARE guidelines: Consensus-based clinical
with hierarchical linear modeling. Behavior Modification, 37, 62– 89. http:// case report guideline development. Journal of Clinical Epidemiology, 67,
dx.doi.org/10.1177/0145445512453734 46 –51. http://dx.doi.org/10.1016/j.jclinepi.2013.08.003
DeCarlo, L. T., & Tryon, W. W. (1993). Estimating and testing autocorrelation Gast, D. L. (2010a). Replication. In D. L. Gast (Ed.), Single subject research
with small samples: A comparison of the C-statistic to a modified estimator. methodology in behavioral sciences (pp. 110 –128). New York, NY: Rout-
Behaviour Research and Therapy, 31, 781–788. http://dx.doi.org/10.1016/ ledge.
0005-7967(93)90009-J Gast, D. L. (Ed.). (2010b). Single subject research methodology in behavioral
Didden, R., Korzilius, H., van Oorsouw, W., & Sturmey, P. (2006). Behavioral sciences (1st ed.). New York, NY: Routledge.
treatment of challenging behaviors in individuals with mild mental retarda- Glasziou, P., Meats, E., Heneghan, C., & Shepperd, S. (2008). What is missing
tion: Meta-analysis of single-subject research. American Journal on Mental from descriptions of treatment in trials and reviews? British Medical Jour-
Retardation, 111, 290 –298. http://dx.doi.org/10.1352/0895-8017(2006) nal, 336, 1472–1474. http://dx.doi.org/10.1136/bmj.39590.732037.47
111[290:BTOCBI]2.0.CO;2 Golder, S., Loke, Y. K., & Bland, M. (2011). Meta-analyses of adverse effects
Dijkers, M., Kropp, G. C., Esper, R. M., Yavuzer, G., Cullen, N., & Bakdalieh, data derived from randomised controlled trials as compared to observational
Y. (2002). Quality of intervention research reporting in medical rehabilita- studies: Methodological overview. PLoS Medicine, 8, e1001026. http://dx
tion journals. American Journal of Physical Medicine & Rehabilitation, 81, .doi.org/10.1371/journal.pmed.1001026
21–33. http://dx.doi.org/10.1097/00002060-200201000-00005 Gorsuch, R. L. (1983). Three methods for analyzing limited time-series (N of
Ducharme, J. M., Davidson, A., & Rushford, N. (2002). Treatment of oppo- 1) data. Behavioral Assessment, 5, 141–154.
sitional behavior in children of parents with brain injury and chronic pain. Gottman, J. M., & Glass, G. V. (1978). Analysis of interrupted time-series
Journal of Emotional and Behavioral Disorders, 10, 241–248. http://dx.doi experiments. In T. R. Kratochwill (Ed.), Single-subject research: Strategies
.org/10.1177/10634266020100040601 for evaluating change (pp. 197–235). New York, NY: Academic Press.
Dwan, K., Altman, D. G., Cresswell, L., Blundell, M., Gamble, C. L., & http://dx.doi.org/10.1016/B978-0-12-425850-1.50011-9
Williamson, P. R. (2011). Comparison of protocols and registry entries to Gravetter, F. J., & Forzano, L. A. B. (2009). Research methods for the
published reports for randomised controlled trials. Cochrane Database of behavioral sciences (3rd ed.). Belmont, CA: Wadsworth.
Systematic Reviews, 1, MR000031. http://dx.doi.org/10.1002/14651858 Guyatt, G., Jaeschke, R., & McGinn, T. (2002). N-of-1 randomized controlled
.mr000031.pub2 trials. In G. Guyatt & D. Rennie (Eds.), Users’ guides to the medical
Edgington, E. S. (1980). Validity of randomization tests for one-subject literature (pp. 275–290). Chicago, IL: American Medical Association.
experiments. Journal of Educational Statistics, 5, 235–251. http://dx.doi Hagermoser Sanetti, L. M., & Kratochwill, T. R. (2007–2008). Treatment
.org/10.2307/1164966 integrity in behavioral consultation: Measurement, promotion, and out-
Edgington, E. S. (1996). Randomized single-subject experimental designs. comes. International Journal of Behavioral and Consultation Therapy, 4,
Behaviour Research and Therapy, 34, 567–574. http://dx.doi.org/10.1016/ 95–114.
0005-7967(96)00012-5 Hagermoser Sanetti, L. M., & Kratochwill, T. R. (Eds.). (2014). Treatment
Edinger, J. D., Wohlgemuth, W. K., Radtke, R. A., Marsh, G. R., & Quillian, integrity: A foundation for evidence-based practice in applied psychology.
R. E. (2001). Cognitive behavioral therapy for treatment of chronic primary Washington, DC: American Psychological Association. http://dx.doi.org/10
insomnia: A randomized controlled trial. Journal of the American Medical .1037/14275-000
Association, 285, 1856 –1864. http://dx.doi.org/10.1001/jama.285.14.1856 Hammond, D., & Gast, D. L. (2010). Descriptive analysis of single subject
Ehlhardt, L. A., Sohlberg, M. M., Glang, A., & Albin, R. (2005). TEACH-M: research designs: 1983–2007. Education and Training in Autism and De-
A pilot study evaluating an instructional sequence for persons with impaired velopmental Disabilities, 45, 187–202. http://www.jstor.org/stable/
memory and executive functions. Brain Injury, 19, 569 –583. http://dx.doi 23879806
.org/10.1080/002699050400013550 Hayes, S. C., Barlow, D. H., & Nelson-Gray, R. O. (1999). The scientist
Feeney, T. J. (2010). Structured flexibility: The use of context-sensitive practitioner: Research and accountability in the age of managed care (2nd
self-regulatory scripts to support young persons with acquired brain injury ed.). Boston, MA: Pearson.
and behavioral difficulties. The Journal of Head Trauma Rehabilitation, 25, Hedges, L. V., Pustejovsky, J. E., & Shadish, W. R. (2013). A standardized
416 – 425. http://dx.doi.org/10.1097/HTR.0b013e3181fbc0a2 mean difference effect size for multiple baseline designs across individuals.
Ferron, J. M., & Levin, J. R. (2014). Single-case permutation and randomiza- Research Synthesis Methods, 4, 324 –341. http://dx.doi.org/10.1002/jrsm
tion statistical tests: Present status, promising new developments. In T. R. .1086
Kratochwill and J. R. Levin (Eds.), Single-case intervention research: Hickey, E., Bourgeois, M., & Olswang, L. (2004). Effects of training volun-
Methodological and statistical advances (pp. 153–183). Washington, DC: teers to converse with nursing home residents with aphasia. Aphasiology, 18,
American Psychological Association. http://dx.doi.org/10.1037/14376-006 625– 637. http://dx.doi.org/10.1080/02687030444000093
28 TATE ET AL.

Higginbotham, D. J., & Bedrosian, J. (1995). Subject selection in AAC Lexchin, J., Bero, L. A., Djulbegovic, B., & Clark, O. (2003). Pharmaceutical
research: Decision points. Augmentative and Alternative Communication, industry sponsorship and research outcome and quality: Systematic review.
11, 11–13. http://dx.doi.org/10.1080/07434619512331277099 British Medical Journal, 326, 1167–1170. http://dx.doi.org/10.1136/bmj
Hoffmann, T. C., Glasziou, P. P., Boutron, I., Milne, R., Perera, R., Moher, D., .326.7400.1167
. . . Michie, S. (2014). Better reporting of interventions: Template for Lloyd, B., & Sherrington, C. (2011). Exercise-based videogames for stroke
Intervention Description and Replication (TIDieR) checklist and guide. rehabilitation at home: A single subject randomised trial. Retrieved from
British Medical Journal, 348, g1687. http://dx.doi.org/10.1136/bmj.g1687 https://www.anzctr.org.au/Trial/Registration/TrialReview.
Holcombe, A., Wolery, M., & Snyder, E. (1994). Effects of two levels of aspx?id⫽343162&isClinicalTrial⫽False
procedural fidelity with constant time delay on children’s learning. Journal Lo, Y.-y., Cooke, N. L., & Starling, A. L. P. (2011). Using a repeated reading
of Behavioral Education, 4, 49 –73. http://dx.doi.org/10.1007/BF01560509 program to improve generalization of oral reading fluency. Education &
Hoogeboom, T. J., Kwakkenbos, L., Rietveld, L., den Broeder, A. A., de Bie, Treatment of Children, 34, 115–140. http://dx.doi.org/10.1353/etc.2011
R. A., & van den Ende, C. H. M. (2012). Feasibility and potential effec- .0007
tiveness of a non-pharmacological multidisciplinary care programme for Loke, Y. K., Price, D., & Herxheimer, A., & the Cochrane Adverse Effects
persons with generalised osteoarthritis: A randomised, multiple-baseline Methods Group. (2007). Systematic reviews of adverse effects: Framework
single-case study. British Medical Journal Open, 2, e001161. http://dx.doi for a structured approach. BMC Medical Research Methodology, 7, 32.
.org/10.1136/bmjopen-2012-001161 http://dx.doi.org/10.1186/1471-2288-7-32
Hopewell, S., Clarke, M., Moher, D., Wager, E., Middleton, P., Altman, D. G., Maggin, D. M., Briesch, A. M., Chafouleas, S. M., Ferguson, T. D., & Clark,
. . . the CONSORT Group. (2008). CONSORT for reporting randomized C. (2014). A comparison of rubrics for identifying empirically supported
controlled trials in journal and conference abstracts: Explanation and elab- practices with single-case research. Journal of Behavioral Education, 23,
oration. PLoS Medicine, 5, e20. http://dx.doi.org/10.1371/journal.pmed 287–311. http://dx.doi.org/10.1007/s10864-013-9187-z
.0050020 Maggin, D. M., Chafouleas, S. M., Goddard, K. M., & Johnson, A. H. (2011).
Horner, R. H., Carr, E. G., Halle, J., McGee, G., Odom, S., & Wolery, M. A systematic evaluation of token economies as a classroom management
(2005). The use of single-subject research to identify evidence-based prac- tool for students with challenging behavior. Journal of School Psychology,
tice in special education. Exceptional Children, 71, 165–179. http://dx.doi 49, 529 –554. http://dx.doi.org/10.1016/j.jsp.2011.05.001
.org/10.1177/001440290507100203 Maggin, D. M., & Odom, S. L. (2014). Evaluating single-case research data for
Ioannidis, J. P., Evans, S. J., Gøtzsche, P. C., O’Neill, R. T., Altman, D. G., systematic review: A commentary for the special issue. Journal of School
Schulz, K., . . . the CONSORT Group. (2004). Better reporting of harms in Psychology, 52, 237–241. http://dx.doi.org/10.1016/j.jsp.2014.01.002
randomized trials: An extension of the CONSORT Statement. Annals of Maggin, D. M., O’Keeffe, B. V., & Johnson, A. H. (2011). A quantitative
Internal Medicine, 141, 781–788. http://dx.doi.org/10.7326/0003-4819-141- synthesis of methodology in the meta-analysis of single-subject research for
10-200411160-00009 students with disabilities: 1985–2009. Exceptionality, 19, 109 –135. http://
Kazdin, A. E. (1982a). Single-case experimental designs in clinical research dx.doi.org/10.1080/09362835.2011.565725
and practice. New Directions for Methodology of Social & Behavioral Manolov, R., Gast, D. L., Perdices, M., & Evans, J. J. (2014). Single-case
Science, 13, 33– 47. experimental designs: Reflections on conduct and analysis. Neuropsycho-
Kazdin, A. E. (1982b). Single-case research designs: Methods for clinical and logical Rehabilitation, 24(3– 4):634 – 660. http://dx.doi.org/10.1080/
applied settings. New York, NY: Oxford University Press. 09602011.2014.903199
Kazdin, A. E. (2011). Single-case research designs: Methods for clinical and Mechling, L. C. (2006). Comparison of the effects of three approaches on the
applied settings. New York, NY: Oxford University Press. frequency of stimulus activations, via a single switch, by students with
Kratochwill, T. R., Hitchcock, J., Horner, R. H., Levin, J. R., Odom, S. L., profound intellectual disabilities. The Journal of Special Education, 40,
Rindskopf, D. M., & Shadish, W. R. (2010). Single-case designs technical 94 –102. http://dx.doi.org/10.1177/00224669060400020501
documentation. Retrieved from http://ies.ed.gov/ncee/wwc/pdf/wwc_scd Mechling, L. C., & Gast, D. L. (2010). Ethical principles and practices. In D. L.
.pdf Gast (Ed.), Single subject research methodology in behavioral sciences (pp.
Kratochwill, T. R., Hitchcock, J. H., Horner, R. H., Levin, J. R., Odom, S. L., 32–56). New York, NY: Routledge Publishers.
Rindskopf, D. M., & Shadish, W. R. (2013). Single-case intervention Mesagno, C., Marchant, D., & Morris, T. (2009). Alleviating choking: The
research design standards. Remedial and Special Education, 34, 26 –38. sounds of distraction. Journal of Applied Sport Psychology, 21, 131–147.
http://dx.doi.org/10.1177/0741932512452794 http://dx.doi.org/10.1080/10413200902795091
Kratochwill, T. R., & Levin, J. R. (2010). Enhancing the scientific credibility Methe, S. A., & Hintze, J. M. (2003). Evaluating teacher modeling as a
of single-case intervention research: Randomization to the rescue. Psycho- strategy to increase student reading behavior. School Psychology Review,
logical Methods, 15, 124 –144. http://dx.doi.org/10.1037/a0017736 32, 617– 623.
Kratochwill, T. R., & Levin, J. R. (2014). Single-case intervention research: Miltenberger, R. G., Flessner, C., Gatheridge, B., Johnson, B., Satterlund, M.,
Methodological and statistical advances. Washington, DC: American Psy- & Egemo, K. (2004). Evaluation of behavioral skills training to prevent gun
chological Association. http://dx.doi.org/10.1037/14376-000 play in children. Journal of Applied Behavior Analysis, 37, 513–516. http://
Kravitz, R., Duan, N., eds, & the DEcIDE Methods Center N-of-1 Guidance dx.doi.org/10.1901/jaba.2004.37-513
Panel. (2014). Design and implementation of N-of-1 trials: A user’s guide. Moher, D., Hopewell, S., Schulz, K. F., Montori, V., Gøtzsche, P. C., De-
AHRQ Publication No. 13(14)-EHC122-EF. Retrieved from www vereaux, P. J., . . . Altman, D. G. (2010a). CONSORT 2010 explanation and
.effectivehealthcare.ahrq.gov/N-1-Trials.cfm elaboration: Updated guidelines for reporting parallel group randomised
Krishef, C. H. (1991). Fundamental approaches to single subject design and trials. British Medical Journal, 340, c869. http://dx.doi.org/10.1136/bmj
analysis. Malabar, FL: Krieger Publishing Company. .c869
Lane, J. D., & Gast, D. L. (2014). Visual analysis in single case experimental Moher, D., Schulz, K. F., Simera, I., & Altman, D. G. (2010b). Guidance for
design studies: Brief review and guidelines. Neuropsychological Rehabili- developers of health research reporting guidelines. PLoS Medicine, 7,
tation, 24, 445– 463. http://dx.doi.org/10.1080/09602011.2013.815636 e1000217. http://dx.doi.org/10.1371/journal.pmed.1000217
Lane, K., Wolery, M., Reichow, B., & Rogers, L. (2007). Describing baseline Montgomery, P., Grant, S., Hopewell, S., Macdonald, G., Moher, D., Michie,
conditions: Suggestions for study reports. Journal of Behavioral Education, S., & Mayo-Wilson, E. (2013). Protocol for CONSORT-SPI: An extension
16, 224 –234. http://dx.doi.org/10.1007/s10864-006-9036-4 for social and psychological interventions. Implementation Science; IS, 8,
Leon, S. A., Rosenbek, J. C., Crucian, G. P., Hieber, B., Holiway, B., 99. http://dx.doi.org/10.1186/1748-5908-8-99
Rodriguez, A. D., . . . Gonzalez-Rothi, L. (2005). Active treatments for National Health and Medical Research Council. (2009). NHMRC Australian
aprosodia secondary to right hemisphere stroke. Journal of Rehabilitation Health Ethics Committee (AHEC) position statement: Monitoring and re-
Research and Development, 42, 93–101. http://dx.doi.org/10.1682/jrrd.2003 porting of safety for clinical trials involving therapeutic products. Canberra,
.12.0182 Australia: Author.
 SCRIBE 2016 EXPLANATION AND ELABORATION 29

OCEBM. (2011). The Oxford 2011 levels of evidence. Retrieved from http:// AAC Augmentative and Alternative Communication, 10, 207–223. http://dx
www.cebm.net/index.aspx?o⫽5653 .doi.org/10.1080/07434619412331276920
Onghena, P., & Edgington, E. S. (2005). Customization of pain treatments: Schulz, K. F., Altman, D. G., Moher, D., & the CONSORT Group. (2010).
Single-case design and analysis. The Clinical Journal of Pain, 21, 56 – 68. CONSORT 2010 Statement: Updated guidelines for reporting parallel group
http://dx.doi.org/10.1097/00002508-200501000-00007 randomised trials. BMC Medicine, 8, 18. http://dx.doi.org/10.1186/1741-
Papanikolaou, P. N., Christidi, G. D., & Ioannidis, J. P. A. (2006). Comparison 7015-8-18
of evidence on harms of medical interventions in randomized and nonran- Schulz, K. F., & Grimes, D. A. (2002). Blinding in randomised trials: Hiding
domized studies. CMAJ Canadian Medical Association Journal, 174, 635– who got what. Lancet, 359, 696 –700. http://dx.doi.org/10.1016/S0140-
641. http://dx.doi.org/10.1503/cmaj.050873 6736(02)07816-9
Parker, R. I., & Brossart, D. F. (2003). Evaluating single-case research data: A Scott, D., Scott, L. M., & Goldwater, B. (1997). A performance improvement
comparison of seven statistical methods. Behavior Therapy, 34, 189 –211. program for an international-level track and field athlete. Journal of Applied
http://dx.doi.org/10.1016/S0005-7894(03)80013-8 Behavior Analysis, 30, 573–575. http://dx.doi.org/10.1901/jaba.1997.30-573
Parker, R. I., & Vannest, K. (2009). An improved effect size for single-case Scruggs, T. E., & Mastropieri, M. A. (2013). PND at 25: Past, present, and
research: Nonoverlap of all pairs. Behavior Therapy, 40, 357–367. http://dx future trends in summarizing single-subject research. Remedial and Special
.doi.org/10.1016/j.beth.2008.10.006 Education, 34, 9 –19. http://dx.doi.org/10.1177/0741932512440730
Parker, R. I., Vannest, K. J., & Brown, L. (2009). The “improvement rate Scruggs, T. E., Mastropieri, M. A., & Casto, G. (1987). The quantitative
difference” for single-case research. Exceptional Children, 75, 135–150. synthesis of single-subject research: Methodology and validation. Remedial
http://ecx.sagepub.com/content/75/2/135.abstract and Special Education, 8, 24 –33. http://dx.doi.org/10.1177/
Parker, R. I., Vannest, K. J., & Davis, J. L. (2014). Non-overlap analysis for 074193258700800206
single-case research. In T. R. Kratochwill & J. R. Levin (Eds.), Single-case Shadish, W. R. (2014). Statistical analyses of single-case designs: The shape of
intervention research: Methodological and statistical advances (pp. 127– things to come. Current Directions in Psychological Science, 23, 139 –146.
151). Washington, DC: American Psychological Association. http://dx.doi http://dx.doi.org/10.1177/0963721414524773
.org/10.1037/14376-005 Shadish, W. R., Hedges, L. V., Pustejovsky, J. E., Boyajian, J. G., Sullivan,
Parker, R. I., Vannest, K. J., Davis, J. L., & Sauber, S. B. (2011). Combining K. J., Andrade, A., & Barrientos, J. L. (2014). A d-statistic for single-case
nonoverlap and trend for single-case research: Tau-U. Behavior Therapy, designs that is equivalent to the usual between-groups d-statistic. Neuropsy-
42, 284 –299. http://dx.doi.org/10.1016/j.beth.2010.08.006 chological Rehabilitation, 24, 528 –553. http://dx.doi.org/10.1080/09602011
Parsonson, B. S., & Baer, D. M. (1986). The graphic analysis of data. In A. .2013.819021
Poling & R. W. Fuqua (Eds.), Research methods in applied behavior Shadish, W. R., Rindskopf, D. M., & Hedges, L. V. (2008). The state of the
analysis: Issues and advances (pp. 157–186). New York, NY: Plenum Press. science in the meta-analysis of single-case experimental designs. Evidence-
http://dx.doi.org/10.1007/978-1-4684-8786-2_8 Based Communication Assessment and Intervention, 2, 188 –196. http://dx
Perdices, M., Schultz, R., Tate, R., McDonald, S., Togher, L., Savage, S., . . . .doi.org/10.1080/17489530802581603
Smith, K. (2006). The evidence base of neuropsychological rehabilitation in Shadish, W. R., & Sullivan, K. J. (2011). Characteristics of single-case designs
acquired brain impairment (ABI): How good is the research? Brain Impair- used to assess intervention effects in 2008. Behavior Research Methods, 43,
ment, 7, 119 –132. http://dx.doi.org/10.1375/brim.7.2.119 971–980. http://dx.doi.org/10.3758/s13428-011-0111-y
Perdices, M., & Tate, R. L. (2009). Single-subject designs as a tool for Shamseer, L., Sampson, M., Bukutu, C., Schmid, C. H., Nikles, J., Tate, R. . . .
evidence-based clinical practice: Are they unrecognised and undervalued? and the CENT group. (2015). CONSORT extension for reporting N-of-1
Neuropsychological Rehabilitation, 19, 904 –927. http://dx.doi.org/10.1080/ trials (CENT) 2015: Explanation and elaboration. British Medical Journal,
09602010903040691 350, h1793. http://dx.doi.org/10.1136/bmj/h1793
Pool, D., Blackmore, A. M., Bear, N., & Valentine, J. (2014). Effects of Sidman, M. (1960). Tactics of scientific research: Evaluating experimental
short-term daily community walk aide use on children with unilateral spastic data in psychology. New York, NY: Basic Books, Inc.
cerebral palsy. Pediatric Physical Therapy, 26, 308 –317. http://dx.doi.org/ Siegel, S., & Castellan, N. J. (1988). Nonparametric statistics for the behav-
10.1097/PEP.0000000000000057 ioral sciences (2nd ed.). New York, NY: McGraw-Hill.
Punja, S., Eslick, I., Duan, N., Vohra, S., & the DEcIDE Methods Center N- Simera, I., Moher, D., Hoey, J., Schulz, K. F., & Altman, D. G. (2010). A
of-1 Guidance Panel. (2014). An ethical framework for N-of-1 trials: Clin- catalogue of reporting guidelines for health research. European Journal of
ical care, quality improvement, or human subjects research? In R. L. Clinical Investigation, 40, 35–53. http://dx.doi.org/10.1111/j.1365-2362
Kravitz, N. Duan, & the DEcIDE Methods Center N-of-1 Guidance Panel .2009.02234.x
(Eds.), Design and implementation of N-of-1 trials: A user’s guide. AHRQ Sismondo, S. (2008a). How pharmaceutical industry funding affects trial
Publication No. 13(14)-EHC122-EF (pp. 13–22). Rockville, MD: Agency outcomes: Causal structures and responses. Social Science & Medicine, 66,
for Healthcare Research and Quality. Retrieved from http://www 1909 –1914. http://dx.doi.org/10.1016/j.socscimed.2008.01.010
.effectivehealthcare.ahrq.gov/N-1-Trials.cfm Sismondo, S. (2008b). Pharmaceutical company funding and its consequences:
Rider, J. D., Wright, H. H., Marshall, R. C., & Page, J. L. (2008). Using A qualitative systematic review. Contemporary Clinical Trials, 29, 109 –
semantic feature analysis to improve contextual discourse in adults with 113. http://dx.doi.org/10.1016/j.cct.2007.08.001
aphasia. American Journal of Speech-Language Pathology, 17, 161–172. Smith, J. D. (2012). Single-case experimental designs: A systematic review of
http://dx.doi.org/10.1044/1058-0360(2008/016) published research and current standards. Psychological Methods, 17, 510 –
Robey, R., Shultz, M., Crawford, A., & Sinner, C. (1999). Single-subject 550. http://dx.doi.org/10.1037/a0029312
clinical-outcome research: Designs, data, effect sizes, and analyses. Apha- Student. (1908). The probable error of a mean. Biometrika, 6, 1–25. http://
siology, 13, 445– 473. http://dx.doi.org/10.1080/026870399402028 www.jstor.org/stable/2331554
Rose, M., & Sussmilch, G. (2008). The effects of semantic and gesture Swaminathan, H., Rogers, H. J., Horner, R. H., Sugai, G., & Smolkowski, K.
treatments on verb retrieval and verb use in aphasia. Aphasiology, 22, (2014). Regression models and effect size measures for single case designs.
691–706. http://dx.doi.org/10.1080/02687030701800800 Neuropsychological Rehabilitation, 24, 554 –571. http://www.tandfonline.
Satir, D. A., Goodman, D. M., Shingleton, R. M., Porcerelli, J. H., Gorman, com/doi/abs/10.1080/09602011.2014.887586
B. S., Pratt, E. M., . . . Thompson-Brenner, H. (2011). Alliance-focused Tate, R. L., Perdices, M., McDonald, S., Togher, L., & Rosenkoetter, U. (2014).
therapy for anorexia nervosa: Integrative relational and behavioral change The design, conduct and report of single-case research: Resources to improve
treatments in a single-case experimental design. Psychotherapy: Theory, the quality of the neurorehabilitation literature. Neuropsychological Rehabilita-
Research, & Practice, 48, 401– 420. http://dx.doi.org/10.1037/a0026216 tion, 24, 315–331. http://dx.doi.org/10.1080/09602011.2013.875043
Schlosser, R. W., & Braun, U. (1994). Efficacy of AAC interventions: Meth- Tate, R. L., Perdices, M., Rosenkoetter, U., Shadish, W., Vohra, S., Barlow,
odologic issues in evaluating behavior change, generalization, and effects. D. H., . . . Wilson, B. (2016). The Single-Case Reporting guideline In
30 TATE ET AL.

BEhavioural interventions (SCRIBE) 2016 statement. Archives of Scientific back pain. Behaviour Research and Therapy, 39, 151–166. http://dx.doi.org/
Psychology, 4, 1–9. 10.1016/S0005-7967(99)00174-6
Tate, R. L., Perdices, M., Rosenkoetter, U., Wakim, D., Godbee, K., Togher, Vohra, S., Shamseer, L., Sampson, M., Bukutu, C., Schmid, C. H., Tate, R., . . .
L., & McDonald, S. (2013a). Manual for the critical appraisal of single-case for the CENT group. (2015). CONSORT extension for reporting N-of-1
reports using the Risk of Bias in N-of-1 Trials (RoBiNT) Scale. Unpublished trials (CENT) 2015 Statement. British Medical Journal, 350, h1738. http://
manuscript, Rehabilitation Studies Unit, Sydney Medical School - Northern, dx.doi.org/10.1136/bmj/h1738
The University of Sydney, Sydney, Australia. Wambaugh, J. L., Mauszycki, S., Cameron, R., Wright, S., & Nessler, C.
Tate, R. L., Perdices, M., Rosenkoetter, U., Wakim, D., Godbee, K., Togher, (2013). Semantic feature analysis: Incorporating typicality treatment and
L., & McDonald, S. (2013b). Revision of a method quality rating scale for mediating strategy training to promote generalization. American Journal of
single-case experimental designs and N-of-1 trials: The 15-item Risk of Bias Speech-Language Pathology, 22, S334 –S369. http://dx.doi.org/10.1044/
in N-of-1 Trials (RoBiNT) Scale. Neuropsychological Rehabilitation, 23, 1058-0360(2013/12-0070)
619 – 638. http://dx.doi.org/10.1080/09602011.2013.824383 Wambaugh, J. L., Nessler, C., & Wright, S. (2013). Modified response elab-
Tate, R. L., Rosenkoetter, U., Sigmundsdottir, L., Wakim, D., Doubleday, J., oration training: Application to procedural discourse and personal recounts.
Togher, L., . . .Perdices, M. (2015). The Risk of Bias in N-of-1 Trials American Journal of Speech-Language Pathology, 22, S409 –S425. http://
(RoBiNT) Scale: An expanded manual for the critical appraisal of single- dx.doi.org/10.1044/1058-0360(2013/12-0063)
case reports. Sydney, Australia: Author. Warren, S. F., Fey, M. E., & Yoder, P. J. (2007). Differential treatment
ter Kuile, M. M., Bulté, I., Weijenborg, P. T., Beekman, A., Melles, R., & intensity research: A missing link to creating optimally effective commu-
Onghena, P. (2009). Therapist-aided exposure for women with lifelong nication interventions. Mental Retardation and Developmental Disabilities
vaginismus: A replicated single-case design. Journal of Consulting and
Research Reviews, 13, 70 –77. http://dx.doi.org/10.1002/mrdd.20139
Clinical Psychology, 77, 149 –159. http://dx.doi.org/10.1037/a0014273
Wenman, R., Bowen, A., Tallis, R. C., Gardener, E., Cross, S., & Niven, D.
Treadwell, K., & Page, T. J. (1996). Functional analysis: Identifying the
(2003). Use of a randomised single case experimental design to evaluate
environmental determinants of severe behavior disorders. The Journal of
therapy for unilateral neglect. Neuropsychological Rehabilitation, 13, 441–
Head Trauma Rehabilitation, 11, 62–74. http://dx.doi.org/10.1097/
459. http://dx.doi.org/10.1080/09602010343000048
00001199-199602000-00008
White, D., Rusch, F., Kazdin, A., & Hartmann, D. (1989). Applications of
Tryon, W. W. (1982). A simplified time-series analysis for evaluating treat-
meta-analysis in individual-subject research. Behavioral Assessment, 11,
ment interventions. Journal of Applied Behavior Analysis, 15, 423– 429.
281–296.
http://dx.doi.org/10.1901/jaba.1982.15-423
White, O. R., & Haring, N. G. (1980). Exceptional teaching. Columbus, OH:
Turner, L., Shamseer, L., Altman, D. G., Weeks, L., Peters, J., Kober, T., . . .
Moher, D. (2012). Consolidated standards of reporting trials (CONSORT) Merrill.
and the completeness of reporting of randomised controlled trials (RCTs) Wilcoxin, F. (1947). Probability tables for individual comparisons by ranking
published in medical journals. Cochrane Database of Systematic Reviews, methods. Biometrics, 3, 119 –122. http://dx.doi.org/10.2307/3001946
11, MR000030. http://systematicreviewsjournal.biomedcentral.com/articles/ Wolery, M. (1994). Procedural fidelity: A reminder of its functions. Journal of
10.1186/2046-4053-1-60 Behavioral Education, 4, 381–386. http://dx.doi.org/10.1007/BF01539539
Valdimarsdóttir, H., Halldórsdóttir, L. Y., & SigurÐardóttir, Z. G. (2010). Wolery, M., Dunlap, G., & Ledford, J. R. (2011). Single-case experimental
Increasing the variety of foods consumed by a picky eater: Generalization of methods: Suggestions for reporting. Journal of Early Intervention, 33,
effects across caregivers and settings. Journal of Applied Behavior Analysis, 103–109. http://dx.doi.org/10.1177/1053815111418235
43, 101–105. http://dx.doi.org/10.1901/jaba.2010.43-101 Wolery, M., & Ezell, H. K. (1993). Subject descriptions and single-subject
Vandenbroucke, J. P. (2006). What is the best evidence for determining harms research. Journal of Learning Disabilities, 26, 642– 647. http://dx.doi.org/
of medical treatment? CMAJ Canadian Medical Association Journal, 174, 10.1177/002221949302601001
645– 646. http://dx.doi.org/10.1503/cmaj.051484 Wood, L., Egger, M., Gluud, L. L., Schulz, K. F., Jüni, P., Altman, D. G., . . .
Vandenbroucke, J. P., von Elm, E., Altman, D. G., Gøtzsche, P. C., Mulrow, Sterne, J. A. C. (2008). Empirical evidence of bias in treatment effect
C. D., Pocock, S. J., . . . the STROBE Initiative. (2007). Strengthening the estimates in controlled trials with different interventions and outcomes:
Reporting of Observational Studies in Epidemiology (STROBE): Explana- Meta-epidemiological study. British Medical Journal, 336, 601– 605. http://
tion and elaboration. PLoS Medicine, 4, e297. dx.doi.org/10.1136/bmj.39465.451748.AD
Vlaeyen, J. W., de Jong, J., Geilen, M., Heuts, P. H., & van Breukelen, G. Wright, H. H., Marshall, R. C., Wilson, K. B., & Page, J. L. (2008). Using a
(2001). Graded exposure in vivo in the treatment of pain-related fear: A written cueing hierarchy to improve verbal naming in aphasia. Aphasiology,
replicated single-case experimental design in four patients with chronic low 22, 522–536. http://dx.doi.org/10.1080/02687030701487905

(Appendix follows)
 SCRIBE 2016 EXPLANATION AND ELABORATION 31

Appendix
Selection of Data Analysis Methods Applicable to SCEDs

For a comprehensive discussion of statistical techniques in SCEDs, Effect Sizes

see Kazdin (2011).
For reviews, see Alresheed, Hott, and Bano (2013); Beretvas and
Visual Analysis Chung (2008); Shadish, Rindskopf, and Hedges (2008):

• Guidelines for systematic analysis (Kratochwill et al., 2010) • Trend analysis effect size (Gorsuch, 1983)
• Software for visual analysis (Bulté & Onghena, 2012) • Improvement rate difference (Parker, Vannest, & Brown, 2009)
• Nonoverlap of all pairs* (Parker & Vannest, 2009)
Quasistatistical Techniques • Percentage of nonoverlapping data* (Scruggs & Mastropieri,
2013; Scruggs, Mastropieri, & Casto, 1987)
• Split-middle trend line (Kazdin, 1982a; White & Haring, 1980) • Tau-U (Parker, Vannest, David & Sauber, 2011)
• Binomial distribution test (Siegel & Castellan, 1988) • Bayesian probability model
• Standard deviation band (Bloom & Fischer, 1982; Krishef, 1991) • Mean-shift and mean-plus trend models (Allison & Gorman,
1993; Center, Skiba, & Casey, 1985)
Time Series Analysis • d Statistic (Hedges, Pustejovsky, & Shadish, 2013; Shadish,
2014)
• C statistic (DeCarlo & Tryon, 1993; Tryon, 1982)
• Multilevel linear modelling (Swaminathan, Rogers, Horner,
• Auto-regressive integrated moving average (Box & Jenkins,
Sugai, & Smolkowski, 2014)
1970; Gottman & Glass, 1978)

“Traditional” Inferential Statistics Tests Randomization Tests (Edgington, 1980, 1996; Ferron &
Onghena, 1996; Ferron & Ware, 1994; Onghena &
Parametric. Edgington, 2005)
• t Test (Student, 1908) * For more information on nonoverlap methods, see Parker, Vannest,
• F Test (Fisher, 1920) and Davis (2014).

Nonparametric.
Received August 21, 2015
• Wilcoxon matched-pairs signed-ranks test (Wilcoxon, 1947) Revision received November 15, 2015
• Friedman two-way analysis of variance (Friedman, 1937) Accepted November 16, 2015 䡲