Development of The Interview and Interrogation Assessment Instrument

Psychology, Public Policy, and Law
© 2022 American Psychological Association 2023, Vol. 29, No. 1, 46–61

ISSN: 1076-8971 https://doi.org/10.1037/law0000366
Development of the Interview and Interrogation Assessment Instrument
Jeffrey Kaplan1, Brian L. Cutler2, Amy-May Leach1, and Joseph Eastwood1

1
Faculty of Social Science and Humanities, University of Ontario Institute of Technology
2
Department of Media Psychology, Fielding Graduate University
Confession evidence factors heavily in judicial decision-making, and courts may call an expert social
scientist to assess the coercive pressures of an interrogation and risk factors for false confession. At
present, there exists no standardized methods for performing this task, and each expert uses their own
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
unstructured professional judgment. To address this lack of standardization, we have developed a

psychological instrument for evaluating videotaped interrogations: the Interview and Interrogation
Assessment Instrument. We begin with a discussion of the benefits of standardized measurement and
proceed to an overview of the conceptualization and initial development of our instrument. In Study
1, we established the bases for the instrument’s items and scoring by surveying expert populations. In
Study 2, we assessed interrater reliability and explain our instrument refinements based on our results.
In Study 3, we examined convergent validity. Social science experts reviewed interrogation videos
and rated the coercive pressures along multiple dimensions. We correlated the expert ratings with our
instrument’s measures. We conclude that the newly developed instrument demonstrates preliminary
reliability and convergent validity and appears to be a promising tool for future research and expert
consultation in contested confession cases.
Keywords: interrogations, false confessions, expert witness, coercion
All confessions are not equal. Suspects’ statements under ques- This approach offers the courts more latitude in examining confes-
tioning often play a pivotal role in the prosecution of the guilty. sion evidence and creates opportunities for psychological experts
Unfortunately, confessions are also a key contributor to the wrong- to testify on coercion and false confessions.
ful conviction of the innocent (Kassin et al., 2010). Because of the As judicial interest in the psychology of interrogations has
weight given to confessions among triers of fact, confessions are grown, so too has the sophistication of the social science research.
often contested in criminal cases. Concerns about interrogators The current research base on interrogations and confessions was
coercing false confessions from criminal suspects are not modern built upon reliable science and is generally accepted in its field
in origin; those within judicial and social science spheres have (Gudjonsson, 2021). The body of literature now encompasses ex-
been raising objections to certain interrogation practices since perimental research and meta-analyses (e.g., Klaver et al., 2008;
the early 20th century (e.g., Borchard, 1932; Munsterberg, 1908; Meissner et al., 2012, 2014), surveys of expert populations (e.g.,
Wickersham Commission, 1931), if not longer (e.g., Bram v. Kassin et al., 2007), observational and archival studies (e.g.,
United States, 1897; Hopt v. Utah, 1884; King v. Warickshall, Kaplan & Cutler, 2021; King & Snook, 2009; Ofshe & Leo,
1783). Over time, the courts have gradually moved beyond merely 1997), and review papers authored by psychologists (e.g., Kassin
examining physical abuse, direct threats, and promises to give et al., 2010; Scherr et al., 2020). It has also been highly informed
more consideration to concepts such as a suspect’s “operating by textbooks and training manuals written from the perspective of
mind” and an “atmosphere of oppression” (R v. Oickle, 2000, p. 1). those who conduct suspect questioning (e.g., Carr, 2015; Inbau
This article was published Online First September 29, 2022. Jeffrey Kaplan served as lead for data curation, formal analysis,
Jeffrey Kaplan https://orcid.org/0000-0002-3681-1214 investigation, methodology, project administration, validation, writing –
This research was presented in a symposium at the American Psychology – original draft, and writing – review & editing; contributed equally to
Law Society 2022 Conference. This work was supported by American conceptualization; and served in a supporting role for funding acquisition.
Psychology – Law Society Grants in Student Aid, Social Sciences and Brian L. Cutler served as lead for funding acquisition, contributed equally
Humanities Research Council of Canada CGS-D Grant 67-2020-1482, and to conceptualization and methodology, and served in a supporting role for
Social Sciences and Humanities Research Council of Canada Insight Grant writing – review & editing. Amy-May Leach served in a supporting role for
430-2016-00605. We note that the first, second, and third authors claim conceptualization and writing – review & editing. Joseph Eastwood served
ownership of the Interview and Interrogation Assessment Instrument. We wish in a supporting role for writing – review & editing.
to acknowledge Christopher Kelly, Alison Redlich, and Richard Leo for Correspondence concerning this article should be addressed to Jeffrey
sharing research materials. The data sets and additional tables associated with Kaplan, Faculty of Social Science and Humanities, University of Ontario
this research are available through the Center for Open Science at https://osf.io/ Institute of Technology, 2000 Simcoe Street North, Oshawa, Ontario L1H
r27uq/?view_only=028d8d1a9c14492095721e8623ce7b75. 7K4, Canada. Email: [email protected]
46
DEVELOPMENT OF THE IIAI 47
et al., 2013; Wicklander-Zulawski & Associates, 2020) and legal Other vulnerabilities may be due to the limitations of attention
and criminological critical reviews (e.g., Leo, 2008; Pepson & and memory and related phenomena such as selective attention
Sharifi, 2010). As demonstrated by Kassin et al.’s (2018) survey (Watts & Weems, 2006) and the availability heuristic (Tversky &
of confession experts, there is a strong degree of agreement Kahneman, 1973). When experts use unstructured methods, some
between experts on risk factors for false confession, although relevant factors may escape their attention or not be given proper
much of this knowledge remains outside of the ken of jurors consideration, an effect that has been noted in medicine, mechan-
(Alceste et al., 2021; Jones & Penrod, 2016; Kaplan et al., 2020). ics, aviation, and many other fields (Gawande, 2010).
The exponential growth in research over the past few decades Other biases may arise due to the nature of North American
has greatly enhanced our understanding of police interrogations legal systems. Experts’ assessments of risk of recidivism and civil
but has also made the prospect of analyzing a given interrogation commitment criteria tend to show a systematic bias, or adversarial
much more complex; each study that uncovers new knowledge allegiance (Murrie et al., 2009), toward the determinations pre-
about coercion and false confessions introduces new factors that ferred by the party that retained them (McAuliff & Arter, 2016;
an expert must be cognizant of and incorporate into their analysis. Murrie et al., 2013; Rufino et al., 2012). Subjective unguided
measures are more susceptible to practitioner bias and adversarial
Important factors may be overlooked and/or influenced by bias. In

this article, we introduce a new comprehensive and comparatively allegiance than highly structured methods of assessment (Gowen-
objective method of analyzing coercion in interrogations and smith et al., 2012; Murrie & Boccaccini, 2015).
confessions. A final limitation of evaluating interrogations and confessions
in the absence of any structured instrument is that the resultant
Current Assessments and Their Limitations assessments lack any firm metrics. If an expert were to be asked
how an interrogation and confession compare to others (a question
In criminal law, expert testimony on confession evidence is typi- that, as consultants, we are routinely asked by defense lawyers
cally proffered under one of two contexts. There exist legal protec- who send us recorded interrogations for review), the only answer
tions against being coerced into making statements against one’s they could proffer would be a qualitative statement based on
will under questioning by law enforcement both at the national others they have reviewed in their careers. Thus, the process of
(Constitution Act, 1982; U.S. Const. amend. II) and international evaluating interrogations is also vulnerable to biases resultant
levels (Melzer, 2018). Criminal court judges may hold suppression from an anchoring effect (Tversky & Kahneman, 1973) wherein
hearings to determine the admissibility of a confession by examining experts may, in part, base their opinions on implicit comparisons
its voluntariness (Miranda v. Arizona, 1966; Schulhofer, 2001; to other interrogations they have encountered rather than the col-
Thomas & Leo, 2009), and an expert opinion may assist the trier of lective knowledge of the social science on interrogations and
law in that process. However, absent some explicit violation of a confessions.
suspect’s rights, criminal court judges are typically hesitant to rule a
confession inadmissible (Thomas & Leo, 2014). It is much more The Interview and Interrogation Assessment Instrument
likely that the court will consider issues of coercion to go to the
weight of the confession evidence, rather than its admissibility, and To address the above-mentioned limitations and to introduce
defer to the trier of fact (Pepson & Sharifi, 2010). In such instances, standardization to the evaluation of interrogations and confessions,
the jury may be informed about coercion or risk factors for false we undertook the development of the Interview and Interrogation
confessions by an expert psychological witness (Costanzo & Leo, Assessment Instrument (IIAI).1 The IIAI was designed to capture
2007; Marion et al., 2019; Soree, 2005). factors that affect the coercive pressure of suspect questioning and
In both contexts, an expert would review the interrogation vid- thereby serve as a measure of the risk of a coerced compliant false
eotape, audiotape, and/or transcript and idiosyncratically keep confession. These include factors that relate to the individual being
note of the factors that they deemed relevant and draw conclusions interrogated, those that relate to the environment of the interroga-
about the interrogation using their own professional judgment tion, and the tactics used during the interrogation. The recording
(Kaplan, 2022). There are numerous potential weaknesses and lim- of interrogations is standard practice in many jurisdictions (Bang
itations to such an approach as there is a collection of well-known et al., 2018), and the framework we developed is designed to func-
cognitive biases that may taint evaluators’ unguided judgments tion as an observational coding structure for use while viewing
(see Krauss & Lieberman, 2007; Marion et al., 2019; Neal & videotaped interrogations (see Figures 1 and 2).
Grisso, 2014). Although these biases have not been investigated Although the primary focus of the instrument’s measures pertains
specifically in interrogation evaluation, they are well documented to factors in accusatory interrogations, we opted to include the word
in other areas of forensic assessment. In the field of risk assess- “interview” in the name of our instrument for several reasons.
ment, for instance, unguided judgments have been repeatedly Many models of suspect questioning such as the Reid technique of
shown to be vulnerable to influence from the representativeness interrogation (Inbau et al., 2013) and the phased interview model
heuristic (Krauss & Sales, 2001; Monahan, 1981), which is to (Carr, 2015) begin with an interview that may shift to an
judge something/someone by how closely it matches a schema
1
(Tversky & Kahneman, 1974, 1986). In the context of interroga- The IIAI was originally named and is discussed as the Coercion
tion evaluation, this could hypothetically manifest as evaluations Assessment Instrument in Kaplan et al. (2019). In that chapter, we
introduced the early conceptualization of the instrument and attempts to
being subtly and subconsciously influenced by their own precon- operationalize coercion. The studies discussed in the current article and the
ceptions of how coercive interrogations unfold or how vulnerable analyses of the instrument’s reliability and validity only appear in the
suspects behave. current article.
48 KAPLAN, CUTLER, LEACH, AND EASTWOOD
Figure 1
Active Coding of Interrogation Tactics Using the IIAI
Note. See the online article for the color version of this figure.
interrogation; knowing what took place in the preinterrogation tactics, our instrument may then be used to demonstrate that a
interview is highly pertinent to evaluating the proceeding interroga- confession was obtained without putting the suspect under signifi-
tion. To test and validate the IIAI, we primarily used videotapes of cant coercive pressures.
interrogations, although many included a preinterrogation interview It should be made explicitly clear that the IIAI was designed to
as well. Second, if nonaccusatory and noncoercive methods are measure coercion in interrogations and not the absolute truth of
used to gain a confession, without the use of any accusatory confessions. A coerced confession may be a true confession (i.e.,
Figure 2
Active Coding of Suspect Behaviors Using the IIAI
Note. See the online article for the color version of this figure.
when the suspect is guilty) or a false confession (when the suspect coercive items of the investigator making statements that they are
is innocent). False confessions also come in various forms. A vol- trying to help the suspect and offering incentives for information.
untary false confession is one in which an innocent individual
offers a confession with little or no prompting (Kassin et al., Evidence Based
2010). A coerced compliant false confession occurs when an inno- The evidence-based domain contains tactics that leverage evi-
cent suspect is persuaded or coerced by investigators to confess dence or perceptions of proof over the suspect in some manner. It
(Kassin et al., 2010). In this research, we focus specifically on includes directly presenting evidence, pointing out contradictions
assessing coercion in the interrogation process. We therefore focus in a suspect’s account, as well as deceptive tactics, such as pre-
on risk factors for coerced compliant false confessions, but the senting false evidence, bluffs/bait questions, and referring to the
IIAI was not designed to assess the veracity of confessions. strength of the evidence without revealing specifics.
The IIAI does not replace an expert witness. It provides a means
of organizing findings and produces metrics related to an interroga- Social and Psychological Manipulation
tion and confession. Interpreting that data requires a strong back-
During an accusatory interrogation, an investigator may employ
ground in the psychology of interrogations. Thus, the IIAI is meant
certain maximization techniques (Kassin & McNall, 1991) that

to aid those who already possess a mastery of the psychology of

emphasize the potential consequences the suspect is facing (Inbau
interrogations and confessions in evaluating confession evidence
et al., 2013; Kassin et al., 2010). Typically, these consequences in
and organizing their opinions. IIAI evaluations contain a record of
some way refer to charges or sentencing; however, on occasion,
what occurred in an interrogation, but the psychology underlying
the negative outcomes that an investigator chooses to emphasize
why the interrogation may have been coercive or why the identified
are those related to social outcomes, psychological suffering, or
occurrences are risk factors for false confession still requires the
even religion. The social and psychological manipulation domain
interpretation of a qualified expert. The IIAI can help an expert keep
contains such tactics, including increasing feelings of guilt, attack-
track of all of the pertinent factors and produce metrics about an
ing a suspect’s character, and demeaning comments.
interrogation that no other methods of evaluation currently offer.
Making sense of those metrics and explaining them to the courts is Consequence Based
the responsibility of the expert witness.
The majority of the maximization techniques accounted for by
Items Design the IIAI are found under the consequence-based domain, which
contains tactics that create punishments for noncooperation and
The IIAI items are divided into tactics, the suspect’s behaviors
emphasize the harms of further noncooperation. The domain
and demeanor, investigator demeanors, the physical context and
includes items such as directly accusing the suspect, making direct
environment, and suspect risk factors. The tactic items of the IIAI,
threats, expressing certainty of conviction and harsh sentencing,
as we chose to define them, generally do not refer to extremely
emphasizing the seriousness of the crime, and directly presenting a
specific behaviors, but rather broader constructs. They may, there-
forced choice between two alternative themes.
fore, manifest in an interrogation in various ways. For instance,
the item “degrading others” refers to an interrogator placing blame Minimization
on the victim or a coaccused, whereas “justifications” refers to
blaming something more abstract, such as a temporary loss of con- The items comprising the minimization domain all imply miti-
trol, desperation, or black out theme. The reason for this design is gated punishment in some way. It includes offering justifications,
that interrogations can vary from one another in a virtually infinite appealing to suspects’ self-interest, and downplaying the serious-
number of ways. The details of the crimes, legal and social con- ness of the offense. These items can range from the investigator
texts, and characteristics of the suspects, victims, and investigators making simple statements to constructing sophisticated themes
may differ from one case to another across countless dimensions. that allow the suspect to excuse, normalize, and/or rationalize their
Using broader heuristic categories, rather than fixed sets of spe- criminal actions. The domain also includes direct promises of leni-
cific behaviors, allows the IIAI the flexibility needed to be applica- ency, a tactic prohibited in most jurisdictions.
ble to a wide variety of investigations. Each tactic, in turn, is Investigator Demeanor
organized into the following domains modeled after those first
identified by Kelly et al. (2013). Aside from the five tactic domains listed above, several other
types of variables are accounted for by the IIAI. The investigator
Rapport Building demeanor domain contains a set of items that are coded in a dichot-
Rapport building contains tactics aimed at establishing a social omous format directly on the IIAI scoring sheet. These are judg-
connection between the investigator and suspect, which is most ments about the investigator’s overall behavior throughout the
likely to occur in the early stages of an interrogation (Kaplan & interrogation, such as their use of aggressive questioning (e.g., highly
Cutler, 2021; Kelly et al., 2016). Engaging in rapport building has accusatory tone, asking questions or employing tactics in a rapid-fire
been demonstrated to increase cooperation and the amount of in- manner where the suspect is not given the opportunity to respond).
formation gathered from criminal suspects (Gabbert et al., 2021), Suspect Behaviors and Demeanor
as well as possibly increasing susceptibility to misleading informa-
tion (Vallano & Schreiber Compo, 2015). The items under this do- There are a small number of suspect behaviors pertinent to the
main are engaging in unrelated conversations and empathizing/ evaluation of the interrogation that are accounted for by the IIAI.
sympathizing with suspects. The domain also includes the more These include behaviors indicative of distress (e.g., crying) or
resistance (e.g., denials and objections), making inculpatory state- Luna (2022) defined coercion as arising when “P (i.e., the coercer)
ments, and legally relevant behaviors (e.g., requesting an attorney). is in a position of power and communicates a conditional offer
The IIAI also keeps track of the suspect’s general demeanor. One of and/or threat to Q (i.e., the coercee) who is in a constrained situa-
five demeanors are possible. These demeanors are conglomerations tion created by P” (p. 7). Essentially, the investigator constrains
and amalgamations of more specific behaviors that the content the suspect’s options by removing denial as a viable course of
experts who were consulted in the development of the IIAI indicated action either by making it seem impossible or tying negative con-
were pertinent to evaluating a suspect’s level of emotional duress sequences to it and then makes a conditional offer that indicates
(see Figure 2). that confessing will lead to more desirable outcomes.
Our definition of coercion, however, differs from that offered
Physical Environment and Context by Luna (2022) in two respects; it does not explicitly address the
Physical environment and context are accounted for on the scor- power differential between suspect and investigator (an insightful
ing sheet and include factors such as isolation before interrogation, connection worthy of consideration going forward), and it includes
holding the interrogation in a small and barren room, and the use the psychological wearing down of a suspect via interrogation-
of a polygraph. It also includes a yes/no item to account for related regulatory decline (Davis & Leo, 2012) that is absent from
Luna’s (2022) definition. The ability to resist the demands of

whether the suspect was read a proper legal caution at the start of
questioning. others requires some powers of concentration and self-regulation
(Knowles & Riner, 2007; Muraven & Baumeister, 2000), which
Suspect Risk Factors an investigator may aim to deplete (Davis & Leo, 2012). A sus-
pect’s ability to critically evaluate information, weigh their
There are nine suspect risk factors accounted for in checklist
options, and resist an investigator’s demands may be diminished
form on the IIAI sheet. These include intellectual disability, men-
through repeated coercive tactics that cause fatigue and emotional
tal illness, sleep deprivation, and intoxication. It is not at the IIAI
duress. Our cognitive resources can be metaphorically thought of
user’s discretion to make these determinations based solely on
as an energy source, which can be drained. Eventually, thinking
what they observed in the videotaped interrogation. Rather, this
and reasoning processes become seriously impaired, the ability to
information is generally found in the case’s discovery package in
withstand any more stress dissolves, and the suspect gives in to the
the form of police reports, hospitalization records, and evaluations
demands of the investigator to confess (Davis & Leo, 2012, 2013).
conducted by psychiatrists and clinical psychologists.
It must be noted that our use of the term “coercion” differs from
legal definitions that center on (in)voluntariness. When a confes-
Conceptualization and Development sion is given, it is viewed by the legal system in dichotomous
The above represents the result of 6 years of research and three terms of voluntary or involuntary, where in the latter no option
separate studies. To explain how we arrived at our current point, existed but for a suspect to capitulate to the demands of the crimi-
we begin with a basic summary of the early concept development nal justice system (Hopt v. Utah, 1884; Philips, 1984). Coercion,
and literature review. Before one can endeavor to measure a con- by our definition, is instead viewed on a continuum, and we do not
struct, that construct needs some tangible operational definition. identify an inflection point at which interrogations cross a defini-
Following a thorough review of the legal and psychological litera- tive boundary into being coercive.
ture, we arrived at an operational definition of coercion in interrog- Our operational definition provided a benchmark to gauge the
ation on which to base our measures: types of variables that would need to be measured in order to
effectively evaluate interrogations. Identifying observable and
Coercion in police interrogation consists of the use of persuasive tech- measurable coercive occurrences was necessary to form the foun-
niques that limit the suspect’s autonomy by manipulating the per- dations of an assessment instrument for evaluating interrogations.
ceived costs and benefits of possible courses of action and/or The selection of items, in turn, necessitated the above operationali-
depleting the suspect’s motivation or ability to resist persuasion and zation of the term “coercion” to serve as the criterion by which
acceding to the investigators’ demands. (Kaplan et al., 2019, p. 6) potential items were selected. This process was cyclical. Due to its
intended applications, in constructing the IIAI and selecting poten-
Part of this definition was borrowed from Leo and Liu (2009), tial items for inclusion, special attention was also paid to those
who had defined coercion as “[that which] removes an individual’s risk factors that have been empirically connected to coerced com-
perception of their freedom to make a meaningful choice during a pliant false confessions such as presenting false evidence (Klaver
police interrogation” (p. 385). Once a suspect waives their right to et al., 2008).
silence and an interrogation begins, they essentially have the two For the purposes of interrogation assessment, we began mining
options of denial or confession; many interrogations are premised the literature for the specific interrogation tactics that could contrib-
on making confession seem the better option. Coercive pressure ute to, or potentially detract from, the coerciveness of an interroga-
may be applied by manipulating the perceived costs and benefits tion. Some of the first sources consulted were those that provided
of confessing or continuing with denials and thereby limiting the overviews of interrogations, confessions, and risk factors for false
suspect’s perceived viable options. Our operational definition confession such as review papers (e.g., Kassin et al., 2010) and
therefore includes choice architecture (manipulating perceived books on interrogation (e.g., Gudjonsson, 2003; Leo, 2008).
options and their pros and cons; Knowles & Linn, 2004). The con- Some of our primary sources of tactic items were past observa-
cept of imposing constraints on choices and contrasting them tional studies; because these items had been used previously in the
against conditional offers was also the basis of Luna’s (2022) evaluation of interrogations, accounting for them within our own
recent definition of coercion in interrogations and plea bargaining. observational framework was considered achievable. We retained
the same phrasing from the original sources wherever possible. for false confession (e.g., Cleary, 2017; Malloy et al., 2014; Redlich,
Leo (1996) and King and Snook (2009) provided a number of 2009). Other items added to this section of the IIAI were temporary
items representing interrogation tactics, most of them related to states such as sleep deprivation (Blagrove, 1996; Frenda et al.,
the Reid technique of interrogation (Inbau et al., 2013). Other 2016).
items were drawn from observational studies conducted by Pearse A systematic review performed by Kelly et al. (2013) was our
and Gudjonsson (1999), Bull and Soukara (2010), and Kelly et al. greatest resource throughout the entire development process. The
(2016). Items were also drawn from relevant case law (e.g., Malin- authors set out to create a taxonomy of interrogation tactics; to this
ski v. New York, 1945). A number of surveys regarding interroga- end, they compiled a list of 800 interrogation tactics, combined the
tion and interrogation tactics were consulted in order to identify redundant tactics, removed those that were not applicable, and di-
coercive tactics for possible inclusion in the IIAI. One such source vided the remaining tactics into categories they referred to as
was a survey conducted by Kassin et al. (2007), which, among domains. Although we have made considerable alterations to the
other things, asked police officers about the tactics they regularly taxonomy as originally envisioned by Kelly et al. (2013), such as
use. Expressing anger at suspects, bluffing about evidence, and moving all of the minimization items under one domain and renam-
leveraging polygraph results as evidence during an interrogation
ing the domains, the general structure of meso-level domains that

were items added to the IIAI from this survey. Other surveys that
they identified has been retained in the IIAI.

we reviewed included those that asked laypeople how they per- At the end of our literature review, we identified 194 separate
ceive coercion in interrogations (e.g., Henkel et al., 2008) and one
items (98 tactics across five domains, 65 suspect behaviors, 18
that surveyed forensic psychological experts (Kassin et al., 2018).
environment and context items, and 13 suspect risk factors) for
It is difficult to deem an interrogation as being coercive without
potential inclusion in the IIAI. We then used item reduction strat-
considering the effect on the target of the coercive action (Lidz
egies (Study 1) and reliability analyses (Study 2) to settle on our
et al., 1997; Nozick, 1969). Items representing suspect behaviors
final set of items and scores for our validity tests (Study 3).
indicative of being under coercive duress were gathered from
observational studies such as King and Snook (2009) and Kelly
et al. (2015) and a meta-analysis by DePaulo et al. (2003). Given Study 1
the intended application of the instrument, other legally relevant Study 1 focused on item reduction and had four research objec-
suspect behaviors (e.g., requesting an attorney) were also included tives. The first was to drop any of the 194 items that lacked coer-
among the instrument’s measures.
cive properties; it would be unwieldly for the user to try and
Another set of variables we endeavored to incorporate into the
account for such a large number, and we suspected that many of
measurement of coercion was the environment and context of the
the items were not relevant to an instrument measuring coercion.
interrogation. The length of police questioning was of particular
Second, we sought to further reduce the number of items by com-
concern considering that known false confessions in the past have
bining those items that were conceptually similar and that experts
often been the product of excessively long interrogations (Blair,
thought were comparable. Third, because some of the items may
2005; Drizin & Leo, 2004; Leo & Ofshe, 1998). As to the physical
setting itself, a very small interrogation room where the investiga- have been vaguely phrased or were operationally unclear, we
tor is standing in between the suspect and the exit may create an wanted experts to indicate which items might cause confusion for
atmosphere that amplifies the coercive nature of the situation future IIAI users or be too ambiguous to judge. Fourth, we wished
(Trainum, 2016). There was also a significant amount of literature to establish that the remaining items did, indeed, have coercive
on the effects of isolating suspects before and during interroga- properties and to estimate the relative coercive magnitude of each
tions. Isolating suspects before an interrogation is common prac- item.
tice (Inbau et al., 2013), which is meant to make criminal suspects
feel cut off from social supports. This may decrease suspects’ Participants and Method
sense of autonomy and alter the power dynamics of the situation
We chose to survey two different groups of experts. The first
in a manner that makes it easier for investigators to cut off possible
group was social science experts (SSEs) specializing in interroga-
courses of action and manipulate what the suspect perceives to be
tions and confessions. The recruitment letter for SSEs in Study 1
their options in the interrogation (Scherr et al., 2020). Based on lit-
erature review of risk factors for false confession items such as requested their participation if they had “research interests and ex-
length, isolation, the size of the room, and the time of day were perience in police interviewing of suspects and witnesses, coer-
added to the IIAI under the category of physical environment and cion, or false confessions.” The reasoning for using their expert
context. knowledge as an empirically based assessment of each item was
Personal characteristics that enhance vulnerability to coercion or that they would be highly familiar with the research literature,
the risk of false confession were also relevant items. Suspect risk they would publish and contribute to that literature, and many of
factors, as defined in the IIAI, are permanent or semipermanent traits them would have applied experience in evaluating interrogations
held by suspects that increase their susceptibility to coercion and as expert witnesses. Thus, they would have the expertise needed to
risk of coerced compliant false confessions. Mental illnesses and in- judge the coerciveness of the list of items we had compiled. Crimi-
tellectual disabilities are risk factors for false confession that have nal justice officials (CJOs) from North America also completed
been demonstrated through the examination of past false confessions the survey and were invited to a semistructured interview. CJO
(e.g., Redlich, 2004, 2007; Redlich et al., 2011; Schatz, 2018) and participants were recruited if they had experience “conducting for-
discussed at length theoretically (Gudjonsson, 2003, 2010). There ensic interrogations or evaluating confession evidence as an officer
has also been a great amount of research on youth as a risk factor of the court.” Thus, the CJO sample was comprised of police
investigators, defense attorneys, and prosecuting attorneys. This Tactics that leveraged perceptions of proof were organized into
provided a sample of experts from a variety of perspectives. the evidence-based domain (n = 11 items; Cronbach’s alpha =
A total of 74 participants completed the survey: 54 SSEs and 20 .762). Tactics that leveraged social and psychological outcomes
CJOs. The average age of participants was 45.77 (SD = 13.94). were organized into the social and psychological manipulation do-
The sample was 37.8% (n = 28) male and 62.2% (n = 46) female. main (n = 10 items; Cronbach’s alpha = .740). Those tactics that
Our participants were 84% (n = 62) White, and 16% (n = 12) iden- created or emphasized punishments for noncooperation were put
tified as another race or as biracial. For most items, the ratings of into the consequence-based domain (n = 17 items; Cronbach’s
SSEs and CJOs did not statistically differ from one another (see alpha = .813). Finally, tactics that implied mitigated punishment
Kaplan et al., 2020). were placed into the minimization domain (n = 5 items; Cron-
The survey was hosted by Qualtrics. It contained all 194 items bach’s alpha = .803).
identified through literature review, which were divided into four We reduced the suspect behaviors from 65 to 10 by first remov-
categories: tactics, suspect behaviors, environment and context, ing eight suspect behaviors due to having neutral properties and
and suspect risk factors. Each item was rated on a 7-point Likert combining two that were similar. We then grouped the remaining
scale. Each scale also had an additional option to indicate an item suspect behavior items into suspect demeanors both conceptually
was too vague or unclear to judge. The text above each item
and by the ratings given by experts. Based on those verbal and

included our operational definition and asked participants, “How physical behaviors that grouped together, we identified five general
coercive would you rate the following occurrences?” Participants suspect demeanors (fearful, anxious/agitated, withdrawn/defeated/
rated each interrogation tactic on a 7-point Likert scale ranging depressed, calm/neutral, and defiant/belligerent). Removed and
from 3 (coercion reducing) to 3 (coercion increasing), with 0 combined items from Study 1 and their associated statistics can be
marked “neutral.” We included negative integers because there found in our online materials (Kaplan, Cutler, Leach, & Eastwood,
was some indication from past research that certain behaviors may 2022; https://osf.io/r27uq).
reduce feelings of coercion (specifically, those that increase feel-
ings of independence and efficacy; Norvoll et al., 2017). Thus, we Conclusion and Discussion
expected that certain items, such as allowing the suspect to control
the conversation, may be perceived by experts as detracting from After analyzing the data from our survey, we had an organized
feelings of duress and an atmosphere of coercion. The scales for collection of items that related to our operational definition of
the suspect risk factors slightly differed. The question text above coercion and had been established by experts as having coercive
these scales read, “Given your own personal understanding of properties. Our survey results indicated that the items were inter-
coercion in police interviews and interrogations, to what degree do nally reliable; what remained was to establish interrater reliability
the following characteristics increase a suspect’s susceptibility to by testing whether they could be reliably accounted for by users.
coercion?” The scale ranged from 1 (not at all) to 7 (to a large In our next study, we collected interview and interrogation videos
extent), again with the option “vague or unclear.” This survey was and recruited research assistants to evaluate them using drafts of
administered as part of Kaplan et al. (2020), which was a second- the IIAI.
ary use of this data. This study was not preregistered. The data
associated with this study are available through the Center for Study 2
Open Science at Kaplan, Cutler, Leach, and Eastwood (2022),
https://osf.io/r27uq. This study was approved by our institution’s The IIAI is powered by V-Note software, shown in Figures 1
research ethics board. and 2. Each tactic and suspect behavior item of the IIAI has a key
that is used to designate that the corresponding item is occurring.
Results and Discussion V-Note keeps track of which items were coded, and when, on its
timeline. These data can be exported into spreadsheet format with
All survey items were first subjected to one-sample t tests. The counts of tactics and suspect behaviors, the length of time they
interrogation tactics, suspect behaviors, and environmental consid- were occurring, when in the video they were occurring, and with
erations were compared to the neutral score of 0; suspect charac- which other tactics and behaviors they closely co-occurred. Envi-
teristics were compared to the lowest score of 1. Those that did ronment and context items and suspect characteristics are recorded
not significantly differ were removed. When 5% or more of partic- manually in yes/no format on the IIAI sheet. In the upper right cor-
ipants labeled an item as vague or unclear, the item was discussed ner, there is a space available for making notes during the assess-
by the research team and removed if the team agreed that the item ment of the interrogation, such as the type of evidence raised or
was too ambiguous to judge from a videorecorded interrogation. the types of justifications offered (see Figure 1). The timelines
We then combined items that were conceptually similar and were from the tactic and suspect behavior coding can then be combined
given similar ratings by the experts that we surveyed. Two items and compared, which improves the efficiency of contamination
relating to the use of a polygraph were also combined, despite hav- analyses by making it clear who brought up what piece of informa-
ing a lower Cronbach’s alpha (see online materials: https://osf.io/ tion first.
r27uq). Ultimately, our analyses of survey results reduced the
number of interrogation tactics items from 98 to 51. Using Kelly Pilot Study
et al.’s (2013) taxonomy as our guide, we then organized those
tactics into the domains. Those tactics that aimed to create a social Undergraduate research assistants were given a brief tutorial on
connection between suspect and investigator were put into the rap- V-Note software, provided with the IIAI coding legend, and then
port building domain (n = 8 items; Cronbach’s alpha = .614). used the IIAI to code interrogation videos along with the first author.
This process entailed independently reviewing and coding interroga- similarities. For example, in pilot testing, there was low agreement
tions using the IIAI, then periodically meeting to compare our for the items of “increase guilt” and “honor/religion.” A great deal
results and assess interrater reliability. Areas in which coding dif- of the responsibility for the lack of interrater agreement lay in the
fered between raters were discussed, steps were taken to resolve any fact that in places where one rater had coded one of these items, a
ambiguities, and we made continual adjustments based on our expe- second rater would code the other. These items had been given
riences applying the instrument. This process took place over similar coercion ratings by the experts surveyed, and all went to
approximately 18 months and involved the coding of 50 interroga- the same general construct. Therefore, for the sake of decreasing
tion videos (i.e., five sets of 10 interrogations). the number of items, these were combined under “increase guilt.”
We used two general units of measurement when testing inter- Items that were removed or combined during pilot testing can be
rater reliability: (a) interrogation and (b) time block. For instance, found in our online materials: Kaplan, Cutler, Leach, and Eastwood
if one user coded 39 pieces of evidence being presented for a given (2022), https://osf.io/r27uq.
interrogation and another coded 31, those two data cases would Based on item clarification, item removals, and item combina-
become a point of comparison for analysis using each interrogations over multiple rounds of coding, reliability steadily improved
tion as the unit of measure in intraclass correlational (ICC 2, k)
to the point that it could be considered good to excellent. Summar-

analyses. Those measures, however, would not necessarily mean ized results can be found in Tables 1 and 2. We then endeavored
that both users had coded the same occurrences. When the to formally establish reliability in the following study.
exported data from two raters are compiled together, they form a
cross-tab matrix of tactics and 90-s time blocks (each column a Method
rater’s coding, each row a 90-s time block). Intraclass correlational
(ICC 2, k) analyses were performed on both the 90-s time block Previous research assistants who had used the IIAI were junior
data and the per-interrogation data to assess the reliability of our undergraduates with little education on interrogations and confes-
coding. ICC coefficients may be interpreted using the following sions who were given relatively little training. Due to their diffi-
guidelines: less than .40 = poor, between .40 and .59 = fair, culties in understanding some of the more nuanced aspects of
between .60 and .74 = good, and between .75 and 1.00 = excellent interrogations, we compiled a 78-page training manual and a train-
(Cicchetti, 1994). Suspect risk factors and environment and con- ing program. We recruited three new research assistants holding
text items were analyzed using Cohen’s j. master’s degrees who were currently completing doctoral studies
The reliability measures from our first rounds of pilot testing and two new research assistants who were in their third year of
were poor. One of the primary reasons for low interrater reliability bachelor studies in psychology, but did not yet hold any degrees or
was that there were too many items to attend to at once. To address certifications, to undergo training.
this, the number of tactic items was reduced from 51 to 36 by The training program took place over 2 days. During the first
dropping extremely low-frequency items. Any item that occurred day of the course, an 85-slide presentation was given to the
less than five times during testing was reviewed by the research research assistants; it took approximately 5 hours to complete.
team and dropped if we agreed that they were not necessary for This first session gave an overview of coercion in police interroga-
measuring coercion in police interrogations. For instance, the tions and covered the suspect behaviors, environment and context
items “eye contact,” “unexpected question,” and “reverse order” items, and suspect risk factors. The second part of the course cov-
were three tactics taken from Vrij et al.’s (2006) overview of ered the coding of interrogation tactics and investigator demean-
detecting deception through increasing cognitive load. These tac- ors. It included a 121-slide presentation and took approximately 6
tics did not occur in any of the interrogations during pilot testing hours to complete. Following the completion of training, the five
and had questionable coercive properties, and thus these items research assistants were given seven interrogation videos totaling
were removed. There were, however, some very rarely occurring 6 hours and 55 min to code, which we then analyzed for interrater
items that were retained due to their clearly coercive properties, reliability. This study was not preregistered. This study was
such as the investigator becoming physically aggressive or telling approved by our institution’s research ethics board.
the suspect they cannot leave until they confess. It would detract
from the instrument to too great a degree if it could not account Results and Discussion
for such occurrences, despite their very low prevalence.
We also combined items that were being mistaken for each We established acceptable levels of interrater reliability for
other, had comparable coercion ratings, and shared conceptual most items and domains, with the exception of rapport building
Table 1
Pilot Study Interrater Reliability by Interrogation
Set Total Rapport Evidence S&P Consequence Minimum Suspect
Set 1 .556, .790 .108, .266 .563, .794 .183, .401 .623, .832 .326, .592
Set 2 .594, .815 .522, .766 .807, .926 .626, .834 .320, .586 .396, .663
Set 3 .764, .907 .308, .572 .787, .917 .921, .972 .393, .660 .631, .837 .498, .665
Set 4 .870, .952 .500, .750 .919, .971 .859, .948 .773, .911 .419, .684 .581, .735
Set 5 .875, .933 .650, .788 .899, .947 .915, .956 .864, .927 .853, .920 .800, .889
Note. Single and averaged ICC 2, k measures. Coding of suspect behaviors began in Set 3 of pilot testing; tactic total calculated independent of suspect
data. S&P is the social and psychological manipulation domain.
Table 2
Pilot Study Interrater Reliability by Time Block
Set 1 .384, .651 .147, .341 .443, .705 .183, .401 .213, .448 .408, .674
Set 2 .329, .596 .295, .556 .397, .664 .626, .834 .378, .645 .297, .558
Set 3 .592, .813 .448, .708 .624, .833 .921, .972 .692, .871 .388, .655 .542, .703
Set 4 .676, .862 .476, .732 .747, .899 .859, .948 .656, .851 .591, .812 .469, .663
Set 5 .741, .851 .685, .813 .779, .876 .915, .956 .713, .832 .723, .839 .618, .764
Note. Single and averaged ICC 2, k measures. Coding of suspect behaviors began in Set 3 of pilot testing; tactic total calculated independent of suspect
data. S&P is the social and psychological manipulation domain.
and suspect behaviors (see Tables 3 and 4). For rapport building, previously assessed using our instrument, and SSEs were blind to
this was often due to disagreements regarding coding rapport these previously calculated scores.
building conversations as one long instance or as many instances. Experts rated the coercive pressures of the interrogation in total
At times, the research assistants also coded very basic procedural and along dimensions reflecting the domains of the IIAI. Their rat-
questions (e.g., name, date of birth, etc.) as rapport-building con- ings were compared to IIAI domain scores, suspect behavior
versation when investigators were merely filling out necessary scores, and total IIAI scores. The purpose of these comparisons
paperwork. Similar issues of coding one long instance or multiple was to demonstrate the convergent validity of the IIAI with the
instances arose when coding the suspect’s behaviors as well. Dis- professional judgment of SSEs currently working in the field of
agreements also arose when research assistants coded denials and interrogation and confession evaluation.
admissions to acts that were not criminal offenses, such as failure
to pay bills or having extramarital affairs. Reliability for social Hypotheses
and psychological manipulation and consequence-based tactics
were also weaker than expected. Again, this was largely due to The first hypothesis was that IIAI tactic measures would positively
whether a tactic was coded as one long instance or multiple instan- correlate with independent SSEs’ evaluations of the total coercive
ces. Some specific items such as “evidence bait,” “downplaying pressure of the interrogations. The second hypothesis was that the
seriousness,” and “alternate question” were recurring sources of experts’ ratings for specific types of coercive pressure (e.g., evidence
disagreement due to being undercoded by some research assistants based, minimization, etc.) would positively correlate with those
but attended to by others. Interrater reliability statistics on each measures from the corresponding IIAI domains. The third hypothesis
individual item are available through our online materials (Kaplan, was that SSEs’ ratings of suspect duress would correlate positively
Cutler, Leach, & Eastwood, 2022; https://osf.io/r27uq). with IIAI measures of suspect behaviors.
The total interrater reliability of the instrument was good, and
research assistants agreed about how they captured the main tac- Method
tics of accusatory interrogations. ICC measures for the evidence-
based domain were excellent, and the minimization measures were Participants
good. There was also little disagreement between research assis-
Because we expected a moderate correlation between SSEs’ rat-
tants for the dichotomously recorded items of the environment and
ings and instrument scores (q = .55), power analysis conducted in
context and suspect risk factors. Other specific domains and tactics
G*Power 3.1 (Faul et al., 2007) indicated that a sample of 35 SSEs
showed poor interrater reliability and will be the focus of future
would be sufficient to statistically demonstrate the relationship. This
revisions and additional interrater reliability research.
was rounded up to 40, but we were ultimately only able to collect 36
responses from qualified SSEs.
Study 3 Participants were recruited through the American Psychology –
Each item was given a weighted coercion score based on an av- Law Society listserv, from personal contacts, and by referral from
erage of the experts’ ratings from Study 1. As tactics are used and other participants. The inclusion criteria for SSEs were that (a)
recorded, the total tactic score and domain scores continues to they held a graduate degree in a relevant area of the social scien-
increase relative to those tactics’ coercion weight. We initiated ces, (b) they had performed expert witness functions pertaining to
Study 3 to examine and establish the validity of those scores. To assessing interrogations (e.g., testifying in court, writing reports or
accomplish this, we asked 36 SSEs to review and evaluate one affidavits) and/or had published peer-reviewed research in the
videotaped interrogation each. All interrogations had been areas of interrogations and confessions, and (c) were based in
Table 3
Study 2 Interrater Reliability by Video and Time Block
Video .944, .985 .285, .615 .970, .992 .673, .892 .704, .905 .647, .880 .487, .851
Block .856, .960 .457, .771 .925, .980 .490, .794 .573, .843 .613, .864 .316, .735
Note. Single and averaged ICC 2, k measures. S&P is the social and psychological manipulation domain.
Table 4 13.9% (n = 5) possession of child pornography and/or soliciting

Average Interrogation Environment and Suspect Characteristic minors for sex; 5.6% (n = 2) sexual assault or rape; 5.6% (n = 2)
Interrater Reliability of All Sets assault and battery or attempted homicide; 5.6% (n = 2) reckless
disregard for life and fleeing police; and 2.5% (n = 1) weapons
Environment and context Suspect risk factors
possession in relation to terrorism. Two of the defendants who
Set j j confessed were later found not guilty at trial. Twenty-one of the
Pilot Set 1 .661 .704 videos had been previously used as research materials in Kaplan
Pilot Set 2 .825 .927 and Cutler (2021); otherwise, to our knowledge, none of the vid-
Pilot Set 3 .671 .769 eos had been used in any previous studies.
Pilot Set 4 .568 1.000
Pilot Set 5 .947 1.000 Procedure
Study 2 .719 .751
At the time of contact, each participant was presented with a
recruitment letter explaining the purpose and the goals of the
North America at the time of the study, due to the fact that the
study. Those who were interested in participating were invited to

IIAI was developed for use in North American legal systems.
click on a Qualtrics link, which forwarded them to the letter of

Each SSE was compensated with a $100 CAD or $75 USD Ama-
informed consent and brief survey to determine their eligibility.
zon gift card for their participation. This study was approved by
After giving consent electronically and being deemed eligible,
our institution’s research ethics board.
Qualtrics randomly chose a video for them to review. After read-
The 36 participants were 52.8% female (n = 19) and 44.4% male
ing the case descriptions and before watching the interrogation
(n = 16), and 2.8% (n = 1) declined to identify their gender. Partici-
videos or proceeding to the evaluations, participants were asked if
pants ranged from 25 to 73 years of age (M = 42.58, SD = 13.83). In
they were already familiar with the case or had conflicts of interest
total, 86.1% (n = 31) of SSEs were White, 8.3% (n = 3) were multi-
(e.g., they were an expert witness on this case). If they indicated
racial, and 5.6% (n = 2) were Black. Of the 36 SSEs, 41.7% (n =
such a conflict existed, they were assigned a different video; this
15) reported never serving as an expert witness but having published
occurred five times. Participants then evaluated the interrogations
peer-reviewed research, 36.1% (n = 13) reported serving in 1–10
using the assessment form described in the Measures section. Par-
criminal cases of disputed confessions, 5.6% (n = 2) reported serv-
ticipants were then thanked, debriefed, and forwarded to a page to
ing in 11–24 cases, 11.1% (n = 4) reported serving in 25–49 cases,
collect their compensation and to volunteer participant referrals
and 5.6% (n = 2) reported serving in 50 or more cases.
for the study.
To ensure that participants had completed the study as required,
IIAI Measures. Two different sets of IIAI measures for each
we took note of when each participant began and completed the
domain were used. The first set of measures of this study was the
study. If the total amount of time was less than the length of the
coercion scores produced by the full instrument. At the time of
interrogation video that they were assigned, their responses would publication, we had analyzed and scored 162 interrogation videos
have been removed from the data set as it was not possible for using the IIAI and observed strong skews and kurtosis for our
them to have completed watching the video before filling out the domains (see Table 5). As such strong skews may interfere with
evaluation. As an attention check, participants were asked “Did inferential statistics using those measures, a common solution is to
the suspect confess?” No SSEs were removed for suspiciously low perform a logarithmic transformation on the data (Berry & Feld-
participation times or failing the attention checks, nor did any man, 1985; Field, 2009); we did so here.
begin the study and not complete participation. The second set of measures was the tactic counts per domain.
Material Although these were not the central measures that we were ulti-
mately seeking to validate, the reason that the tactic counts were
The study was hosted by Qualtrics, and participation took place compared to the experts’ ratings was the fact that the IIAI was cur-
completely online. Qualtrics randomly assigned each participant to rently in use as a guided professional judgment tool that did not
a video by providing a link to a Google Drive folder that contained reference any psychometric scoring. Therefore, validating the tac-
a videotaped interrogation and a brief description of the case. After tic counts as measures of coercion could be taken as a measure of
reviewing those materials, participants returned to Qualtrics and convergent validity between SSEs’ ratings and the IIAI as it was
completed their evaluations via that online platform. being used.
The 36 videos used in this study ranged from 26 to 138 min (M = Expert Ratings. Participants were provided a series of Likert
71.85, SD = 26.3) and were of interrogations that took place in Can- scales on which to rate the coercive pressures of the interrogation.
ada or the United States between 2003 and 2018. Interrogations Response options ranged from 1 (not at all) to 10 (to a very large
were chosen if they took place in North America, the suspects extent), with a midpoint of 5 (to a moderate extent). The first ques-
were 18 years of age or older, and the videos contained the inter- tion gave our operational definition of coercion and asked partici-
rogation in its entirety. The suspects in the videos ranged in age pants to evaluate the extent to which the interrogation was
from 18 to 71 (M = 37, SD = 12.89); 89% of them were male (n = coercive. There were also five questions that reflected the domains
32) and 11% (n = 4) were female. The majority of the suspects of the IIAI. Participants were asked the extent to which the investi-
(83.3%; n = 30) were White, 13.9% (n = 5) were Black, and 2.8% gator(s) (a) engaged in rapport building with the suspect; (b) lever-
(n = 1) were Latinx. All interrogations pertained to offences that aged evidence against the suspect; (c) attempted to emotionally
could be charged at the indictable/felony level: 52.8% (n = 19) provoke or manipulate the suspect; (d) engaged in maximization
homicide; 13.9% (n = 5) grand theft, grand larceny, or fraud; by making direct accusations, dismissing denials and objections,
Table 5
IIAI Scores Descriptive Statistics
Pretransformation Posttransformation
Measure M SD Skew SE skew Kurtosis SE kurtosis M SD Skew SE skew Kurtosis SE kurtosis
Rapport building 3.85 4.51 1.98 .191 4.40 .379 1.22 .84 .21 .191 .82 .379
Evidence based 18.97 23.78 2.27 .191 6.05 .379 2.36 1.18 .08 .191 .78 .379
S&P manipulation 9.32 19.27 3.58 .191 14.56 .379 1.31 1.37 .65 .191 .69 .379
Consequence 39.32 57.18 3.29 .191 14.46 .379 2.85 1.46 .39 .191 .79 .379
Minimization 13.88 20.56 3.19 .191 12.59 .379 1.99 1.23 .01 .191 .78 .379
Suspect behaviors 35.41 41.76 3.45 .191 16.31 .379 3.94 1.14 .49 .191 .45 .379
Total 79.95 108.18 3.03 .191 12.63 .379 4.53 1.56 .93 .191 1.46 .379
Note. N = 162. S&P is the social and psychological manipulation domain.
or emphasizing the certainty of guilt or conviction; and (e) made Discussion

appeals to the suspect’s self-interest, offered excuses or justifica-

tions, or blamed others for the offense. Finally, participants were Our correlation analyses of IIAI scores with SSEs’ ratings
asked to what extent the suspect appeared frightened, anxious, or ranged from strong to weak and nonsignificant. The instrument’s
otherwise under duress. This study was not preregistered. The data measures of rapport building clearly differ from that of experts in
associated with this study are available through the Center for the field, as did those of the suspects’ level of duress. Further
Open Science at Kaplan, Cutler, Leach, Marion, et al. (2022), refinement and study are required to validate these aspects of the
https://osf.io/r27uq. IIAI. On the other hand, the IIAI appears to tap into the coercive
use of evidence, the leveraging of consequences, and minimization
as they are currently conceptualized by SSEs. The same pattern of
Results
results was found for the raw tactic counts. Thus, the IIAI appears
The IIAI measures were treated as the dependent variables in a to be proficient at capturing the basic components of accusatory
series of bivariate correlational analyses with the experts’ ratings. interrogations: confront with evidence and perceptions of proof,
The IIAI measures of rapport building for a given interrogation leverage consequences, and allow a way to save face by employ-
were paired with an expert’s rating of that interrogation’s level of ing minimization techniques (Inbau et al., 2013).
rapport building, and so on for all of the domains.
We first correlated SSEs’ total coercion ratings with those inter- General Discussion
rogations’ transformed total IIAI tactic scores. The correlation was
moderate, r = .41, F(1, 34) = 6.735, p = .014. We then conducted Over the course of three studies, we have laid a solid empirical
bivariate correlational analyses for the SSEs’ ratings for each do- framework for the first standardized psychological instrument for
main against IIAI domain scores. The correlational analysis meas- measuring coercion in forensic interrogations. Providing experts
uring the relationship between the transformed rapport-building with a firmly established collection of items to remain aware of
scores and SSEs’ ratings was significant but weak, r = .33, F(1, can ameliorate selective attention by ensuring that experts do not
34) = 4.282, p = .046. We found strong correlations between overlook any important factors when performing extremely com-
experts’ ratings of evidence-based, r = .58, F(1, 34) = 17.376, p , plex tasks (Gawande, 2010), such as evaluating an interrogation
.001, social and psychological manipulation, r = .66, F(1, 34) = for coercion and risk factors for false confession. By providing a
25.727, p , .001, consequence-based, r = .64, F(1, 34) = 24.128, checklist of items with standard definitions, it may also reduce
p , .001, and minimization, r = .48, F(1, 34) = 9.936, p = .003, some of the cognitive biases inherent in socially based judgments.
IIAI scores. Finally, suspect duress scores correlated weakly with Take, for example, the simple occurrence of an investigator accus-
SSEs’ ratings of suspect duress, r = .30, F(1, 35) = 3.499, p = ing a suspect of lying; this item is referred to as “information
.013. These results are summarized in Table 6.
We then conducted a series of bivariate correlational analyses Table 6
comparing IIAI tactic counts for each interrogation to experts’ rat- Study 3 Tactic Counts and IIAI Scores Correlated With Experts’
ings of those interrogations. The total tactic counts showed a mod- Ratings
erate relationship with expert’s ratings of overall coercion, r = .41, Tactic counts IIAI scores
F(1, 34) = 6.81, p = .013. Correlations were strong for evidence-
based tactic counts, r = .51, F(1, 34) = 12.233, p = .001, social and Measure r p r p
psychological manipulation, r = .57, F(1, 34) = 16.048, p , .001, Rapport building .15 .385 .33 .046
consequence-based tactic counts, r = .73, F(1, 34) = 38.359, p , Evidence based .51 .001 .58 ,.001
S&P manipulation .57 ,.001 .66 ,.001
.001, and minimization, r = .53, F(1, 34) = 12.967, p = .001. For Consequences .73 ,.001 .64 ,.001
rapport and relationship building, the relationship was nonsignifi- Minimization .53 .001 .48 .003
cant, r = .15, F(1, 34) = .733, p = .385. The relationship between Suspect duress .29 .079 .30 .035
suspect behavior counts and SSEs’ ratings of duress was nonsigni- Total .41 .013 .41 .014
ficant, r = .29, F(1, 34) = 3.282, p = .079. Note. N = 36. S&P is the social and psychological manipulation domain.
derogation” and is clearly defined and operationalized within our unlikely as all psychological instruments show at least some varia-
instrument. From the IIAI user’s perspective, it does not matter tion between users (Blais et al., 2017; Edens et al., 2016). The
whether they believe that the suspect was lying or if they felt that instrument is still vulnerable to human error, and users may not
the investigator was justified in making that accusation. The user’s always be attentive to all tactics. Each trained IIAI user may have
feelings toward the criminal offense and the imperatives of those their own idiosyncrasies, and the same interrogation evaluated by
who retained them should not interfere with the ability to recog- two experts will invariably show some differences.
nize and code the tactic. Organizing findings based on an empiri- Rapport building was a weak spot in both Studies 2 and 3.
cally established framework, as opposed to subjective clinical Unlike concepts such as minimization and maximization found in
judgment, may therefore serve to reduce biases and adversarial al- accusatory interrogations, the measurement of rapport building has
legiance as it has in the field of risk assessment (Gowensmith been vague and elusive in the social sciences (Gabbert et al.,
et al., 2012; Murrie & Boccaccini, 2015). 2021). In our studies, the construct of rapport building was not so
We also developed a standardized sample of 162 coded inter- easily captured in a set of discretely coded items; this method of
rogations.2 The videotapes used to create this sample pertain to measurement does not take into account nonverbal techniques or
interrogations that took place in North America between 1998 and concepts such as mutual attentiveness (showing caring interest),
2020, ranged from 8 to 491 min (M = 73.38, SD = 67.31), and positivity (positive affect), and coordination (shared understand-
included a variety of felony/indictable offenses. The file contains ing; Tickle-Degnen & Rosenthal, 1990) that experts may have
116 variables and is available to interested researchers through the been attentive to in making their ratings. In practice, this means
Center for Open Science (Kaplan, Rumschick, et al., 2022, https:// the IIAI may not be fully suited to capturing social pressures that
osf.io/r27uq). Newer versions will be periodically added as the may be present in rapport-based methods of questioning, such as
sample continues to grow. Allowing experts to compare data from PEACE (Centrex, 2004). Making refinements to our measures of
a given interrogation to a much larger standardized sample of rapport building will be of focus for future revisions of the IIAI.
interrogations may help reduce the anchoring effect and assist There are also several important expert witness functions that
experts in making comparisons about the coerciveness of an inter- are not aided by the IIAI. It is not a method of evaluating an inves-
rogation to norms. Thus, distilling the past 40 years of research tigator’s proficiency at conducting investigative interviews. The
into a systemized instrument for the first time allows expert wit- IIAI is solely for evaluating coercion in interviews and interroga-
nesses to provide quantifications of coercive pressure based on the tions and does not provide any framework for evaluating the larger
aggregate knowledge of the social science on interrogations and investigation or police tunnel vision. It may assist in confession
confessions. contamination analysis by marking on its timeline when investiga-
In addition to guiding expert witnesses’ judgments, the IIAI will tors raise evidence, but it is not a tool specifically meant for evalu-
be useful in future research for coding interrogations by providing ating the reliability of the information gained from interviews and
common sets of metrics and measures by which to compare results. interrogations.
Users, however, should be mindful of the limitation that interrater Study 3 employed a relatively small number of participants who
reliability varied among the domains. The IIAI produces measures may not fully reflect the state of the social science knowledge on
for 90 different variables relating to the interrogation coded, includ- coercion and interrogations, particularly considering that nearly
ing an exact record of interrogation tactics used and suspect behav- half had never served as expert witnesses. Given the small sample
iors observed. These measures are spread out across numerous size, we also opted not to perform adjustments of our alpha levels;
domains, allowing for the quantification of the amount of coercive we were interested in the strength of the relationships and believed
pressure being applied and how it was being applied. The various it would be overly stringent to apply such a correction given the
measures of the IIAI may be analyzed by themselves, in relation to relatively small sample size.
one another, or in relation to other considerations such as the loca- Our instrument was designed to capture coercive pressure in inter-
tion, the suspects’ demographics (e.g., gender, race), the crime(s) rogations and thereby serve as a measure of the risk of coerced com-
under investigation, etc. pliant false confessions. In Study 3, we established a basic level of
convergent validity with how experts currently working in the crimi-
Limitations nal justice system conceptualize coercive pressure, but we have not
established criterion validity with the outcome of a suspect making a
Although during pilot testing, we established good-to-excellent false confession. With a sample of videotaped interrogations resulting
levels of interrater reliability for all of our domains, in Study 2, in known false confessions, an receiver operating characteristic anal-
there was considerable disagreement for rapport building, social ysis could potentially be performed to demonstrate that those inter-
and psychological manipulation, and suspect behaviors. Some rogations with higher scores are more likely to result in false
individual items such as evidence bait, past acts, and the alternate confessions, and this represents an area of future study. Not having
question also showed low interrater reliability in Study 2. Part of establishing criterion validity does not represent a fatal flaw since an
the reason for this discrepancy between the pilot study and Study expert may not directly opine as to the ultimate question of a confes-
2 may be the fact that the research assistants from the pilot study sion’s veracity or the guilt of a defendant (Marion et al., 2019). Fur-
had much more experience using the instrument by the end of the ther, the majority of the items included in the instrument such
five sets than did the research assistants recruited for Study 2. as explicit threats, promises, minimization and maximization
Therefore, the practice coding of at least five interrogation videos
will be added to the training, and the interrater reliability of the 2
Interrogation data that came from cases in which the authors were
IIAI will be better established through future study. Regardless of retained as expert witnesses have been removed from the publicly available
improvements made, completely perfect interrater reliability is file.
techniques, etc. have been independently confirmed to be risk factors factors are present is not at the IIAI user’s discretion based solely
for false confession through experimental research, archival research, on what they observed in the videotaped interrogation.
and meta-analyses (Drizin & Leo, 2004; Leo & Ofshe, 1998; Meiss- The IIAI was developed in North America by Canadians and
ner et al., 2012, 2014). Americans using a sample of interrogation videos that all came
In Study 1, we removed eight tactics due to experts rating them from North American law enforcement agencies. The basic princi-
as not having coercive properties. While certain tactics such as ples and the items reflecting those principles should be mostly uni-
depersonalizing the situation may not be coercive when used in versal because accusatory approaches are used by a variety of law
isolation, when paired with other tactics (such as a direct accusa- enforcement agencies around the globe. In addition to being
tion), they may have a coercive impact. In such cases where a co- widely used in North America, accusatory and hybrid methods of
ercive tactic is augmented by some other occurrence not directly interrogation, which are premised on choice architecture, are
captured by the IIAI, it may be mentioned in the evaluation, but it taught to and employed by police internationally including in
would not impact the instrument’s scoring. Japan (Wachi & Lamb, 2019), Korea (Reid Institute, 2015), South
The IIAI tactic scores as derived from the results of Study 1 and Africa (Starr, 2019), and across the Middle East (Inbau et al.,
used in Study 3 are not perfect measures. Surveying experts was a 2013) and Latin America (Wicklander-Zulawski & Associates,
proficient method of validating the items that had coercive proper- 2020). However, the instrument has yet to be tested on interroga-
ties and discarding those that did not, but it may not have been an tions from outside of North American jurisdictions and should not
appropriate method to determine the coercive magnitude of each be used to do so until its reliability and validity can be established
item. Ultimately, science is not a democracy where a poll is taken using international samples.
and an averaged middle ground of opinions is accepted as truth,
and our entire basis of the scoring matrix may be predicated on Conclusion
shaky ground. The SSEs’ ratings from Study 3 are also likely less
than perfect reflections of coercive pressure. The experts’ ratings Over the course of three studies, we have laid the empirical
may have been affected by all of the biases and shortcomings of foundations for the IIAI. In Study 1, we were able to greatly
unguided judgments discussed above. The anchoring effect in par- reduce a large number of items down to a workable set of items
ticular may have played a large role in that the IIAI measures were pertinent to the assessment of coercion in police interrogations
produced by a standardized instrument, whereas the expert ratings that were internally reliable. After making significant edits and
were produced not just by unguided judgment but the unguided refinements during pilot testing, at the end of that process, we had
judgment of 36 different individuals, each theoretically with their developed a usable instrument, and we then tested its interrater
own anchoring points. If each of the 36 videos had been reviewed reliability in Study 2. Finally, in Study 3, we established a basic
and judged relatively to one another by a single expert, then the level of convergent validity between most of our measures and the
correlations may have been different. evaluations of coercive pressure of the forensic social science
Throughout the article to this point, we have referred to the IIAI community. Further refinements and testing are needed in some
as an instrument for measuring coercion and risk factors for areas, as discussed in the Limitations section, but the basic frame-
coerced compliant false confession, yet Studies 1 and 3 only asked work that we developed has largely stood up to empirical scrutiny.
participants to rate the former. We maintain that the IIAI will still The field of forensic psychology has been moving toward
assist the courts in adjudicating disputed confessions decisions. In greater standardization for decades (Andrews et al., 2006), and our
order for a confession to constitute admissible inculpatory evi- instrument is a natural step in this process. Equipping forensic
dence, it must be made voluntarily (R v. Piche, 1971; Smith et al., social scientists and expert witnesses with a measurement of coer-
2012). By definition, a coerced statement is not voluntary, and cion based on psychological standards and relevant to criminal
therefore knowing that the items had been rated as coercive by and procedural law will advance both the state of forensic social
experts is useful knowledge for the trier of law in determining science and the effectiveness of the criminal justice system. Our
admissibility. Knowledge about the amount of coercion that eli- instrument is meant to standardize how experts evaluate interroga-
cited a confession is also useful information to jurors in weighing tions in the criminal justice system and provide more comprehen-
veracity; the link between coercive interrogation practices and sive and objective analyses of confession evidence.
false confession is so strong and direct (see Drizin & Leo, 2004;
Kassin et al., 2010; Scherr et al., 2020) that evaluating the amount
of psychological coercion behind a statement is one of the main
References
criteria for evaluating the weight of confession evidence (Kassin, Alceste, F., Luke, T., Redlich, A., Hellgren, J., Amrom, A., & Kassin, S.
1997). (2021). The psychology of confessions: A comparison of expert and lay
The IIAI takes into account suspect risk factors, such as mental opinions. Applied Cognitive Psychology, 35(1), 39–51. https://doi.org/
illness and intellectual disability, but it is not a psychometric 10.1002/acp.3735
instrument for assessing a suspect’s mental capacity. Such diagno- Andrews, D. A., Bonta, J., & Wormith, J. S. (2006). The recent past and
near future of risk and/or need assessment. Crime and Delinquency,
ses must be made by psychiatrists or clinical psychologists; their
52(1), 7–27. https://doi.org/10.1177/0011128705281756
findings may then be incorporated into the IIAI evaluation. Deter-
Bang, B. L., Stanton, D., Hemmens, C., & Stohr, M. K. (2018). Police re-
minations of personal risk factors may be informed by materials cording of custodial interrogations: A state-by-state legal inquiry. Inter-
provided by retaining attorneys such as clinician’s reports, Minne- national Journal of Police Science & Management, 20(1), 3–18. https://
sota Multiphasic Personality Inventory or Wechsler Intelligence doi.org/10.1177/1461355717750172
Scale measures, past histories of psychiatric hospitalization, etc., Berry, W. D., & Feldman, S. (1985). Multiple regression in practice. Sage.
but making the determination as to whether those personal risk https://doi.org/10.4135/9781412985208
Blagrove, M. (1996). Effects of length of sleep deprivation on interrogative Gabbert, F., Hope, L., Luther, K., Wright, G., Ng, M., & Oxburgh, G.
suggestibility. Journal of Experimental Psychology: Applied, 2(1), 48–59. (2021). Exploring the use of rapport in professional information-gathering
https://doi.org/10.1037/1076-898X.2.1.48 contexts by systematically mapping the evidence base. Applied Cognitive
Blair, J. (2005). A test of the unusual false confession perspective using Psychology, 35(2), 329–341. https://doi.org/10.1002/acp.3762
cases of proven false confessions. Criminal Law Bulletin, 41, 127–144. Gawande, A. (2010). The checklist manifesto. Penguin Books India.
Blais, J., Forth, A. E., & Hare, R. D. (2017). Examining the interrater reli- Gowensmith, W. N., Murrie, D. C., & Boccaccini, M. T. (2012). Field reli-
ability of the Hare Psychopathy Checklist-Revised across a large sample ability of competence to stand trial opinions: How often do evaluators
of trained raters. Psychological Assessment, 29(6), 762–775. https://doi agree, and what do judges decide when evaluators disagree? Law and
.org/10.1037/pas0000455 Human Behavior, 36(2) 130–139. https://doi.org/10.1037/h0093958
Borchard, E. M. (1932). Convicting the innocent. Garden City Publishing Gudjonsson, G. H. (2003). The psychology of interrogations and confes-
Company. sions. Wiley.
Bram v. United States 168 U.S. 532 (1897). https://supreme.justia.com/ Gudjonsson, G. H. (2010). Psychological vulnerabilities during police
cases/federal/us/168/532/ interviews. Why are they important? Legal and Criminological Psychol-
Bull, R., & Soukara, S. (2010). Four studies of what really happens in ogy, 15(2), 161–175. https://doi.org/10.1348/135532510X500064
police interviews. In G. D. Lassiter & C. A. Meissner (Eds.), Police Gudjonsson, G. H. (2021). The science-based pathways to understanding
interrogations and false confessions: Current research, practice, and false confessions and wrongful convictions. Frontiers in Psychology,
policy recommendations (pp. 81–95). American Psychological Associa- 12, Article 633936. https://doi.org/10.3389/fpsyg.2021.633936
tion. https://doi.org/10.1037/12085-005 Henkel, L. A., Coffman, K. A., & Dailey, E. M. (2008). A survey of peo-
Constitution Act. (1982). Part I: Canadian charter of rights and freedoms. ple’s attitudes and beliefs about false confessions. Behavioral Sciences
Retrieved August 7, 2022, from http://laws-lois.justice.gc.ca/eng/const/ & the Law, 26(5), 555–584. https://doi.org/10.1002/bsl.826
page-15.html Hopt v. Utah 110 U.S. 574 (1884).
Carr, D. (2015). The RCMP Phased Interview Model for Suspects: Gener- Inbau, F., Reid, J., Buckley, J., & Jayne, B. (2013). Criminal interrogation
ating information through dialogue. Royal Canadian Mounted Police. and confessions (5th ed.). Jones & Bartlett Learning.
Centrex. (2004). Practical guide to investigative interviewing. Central Jones, A. M., & Penrod, S. (2016). Can expert testimony sensitize jurors to
Police Training and Development Authority. coercive interrogation tactics? Journal of Forensic Psychology Practice,
Cicchetti, D. V. (1994). Guidelines, criteria, and rules of thumb for evalu- 16(5), 393–409. https://doi.org/10.1080/15228932.2016.1232029
ating normed and standardized assessment instruments in psychology. Kaplan, J. (2022). Expert testimony on disputed confessions: Bridging
Psychological Assessment, 6(4), 284–290. https://doi.org/10.1037/1040- theory and practice. Faculty of Social Science and Humanities, Ontario
3590.6.4.284 Tech University [Manuscript in preparation].
Cleary, H. M. D. (2017). Applying the lessons of developmental psychology Kaplan, J., & Cutler, B. L. (2021). Co-occurrences among interrogation
to the study of juvenile interrogations: New directions for research, pol- tactics in actual criminal investigations. Psychology, Crime & Law,
icy, and practice. Psychology, Public Policy, and Law, 23(1), 118–130. 28(1), 1–19. https://doi.org/10.1080/1068316X.2021.1876051
https://doi.org/10.1037/law0000120 Kaplan, J., Cutler, B. L., Leach, A. M., & Eastwood, J. (2022). Develop-
Costanzo, M., & Leo, R. A. (2007). Research and expert testimony on ment of the Interview and Interrogation Assessment Instrument study
interrogations and confessions. In M. Costanzo (Ed.), Expert psycholog- three validation study data [Data set]. Center for Open Science. https://
ical testimony for the courts (pp. 69–98). Psychology Press. osf.io/kcqp2?view_only=028d8d1a9c14492095721e8623ce7b75
Davis, D., & Leo, R. A. (2012). Interrogation-related regulatory decline: Kaplan, J., Cutler, B. L., Leach, A. M., Eastwood, J., & Marion, S. (2019).
Ego depletion, failures of self- regulation, and the decision to confess. Evaluating coercion in suspect interviews and interrogations. In B.
Psychology, Public Policy, and Law, 18(4), 673–704. https://doi.org/10 Bornstein & M. Miller’s (Eds.) Advances in psychology and law (pp. 1–40).
.1037/a0027367 Springer. https://doi.org/10.1007/978-3-030-11042-0_1
Davis, D., & Leo, R. A. (2013). Acute suggestibility in police interroga- Kaplan, J., Cutler, B. L., Leach, A. M., Marion, S., & Eastwood, J. (2020).
tion: Self-regulation failure as a primary mechanism of vulnerability. In Perceptions of coercion in interrogation: Comparing expert and lay opin-
A. M. Ridley, F. Gabbert, & D. J. La Rooy (Eds.), Suggestibility in legal ions. Psychology, Crime & Law, 26(4), 384–401. https://doi.org/10
contexts: Psychological research and forensic implications (pp. 171– .1080/1068316X.2019.1669597
195). Wiley-Blackwell. https://doi.org/10.1002/9781118432907.ch9 Kaplan, J., Cutler, B. L., Leach, A. M., Marion, S., & Eastwood, J. (2022).
DePaulo, B. M., Lindsay, J. J., Malone, B. E., Muhlenbruck, L., Charlton, Development of the Interview and Interrogation Assessment Instrument
K., & Cooper, H. (2003). Cues to deception. Psychological Bulletin, study one survey data [Data set]. Center for Open Science. https://osf.io/
129(1), 74–118. https://doi.org/10.1037/0033-2909.129.1.74 n6hgw?view_only=028d8d1a9c14492095721e8623ce7b75
Drizin, S., & Leo, R. (2004). The problem of false confessions in a post- Kaplan, J., Rumschik, D., Woolridge, L. R., Snow, M. D., & Rodriguez
DNA world. North Carolina Law Review, 82, 891–1007. Steen, L. (2022). Interview and Interrogation Assessment Instrument
Edens, J. F., Penson, B. N., Ruchensky, J. R., Cox, J., & Smith, S. T. standardized sample [Data set]. Center for Open Science. https://osf.io/
(2016). Interrater reliability of Violence Risk Appraisal Guide scores kchfq/?view_only=028d8d1a9c14492095721e8623ce7b75
provided in Canadian criminal proceedings. Psychological Assessment, Kassin, S. M. (1997). The psychology of confession evidence. American
28(12), 1543–1549. https://doi.org/10.1037/pas0000278 Psychologist, 52(3), 221–233. https://doi.org/10.1037/0003-066X.52.3
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A .221
flexible statistical power analysis program for the social, behavioral, and Kassin, S. M., Drizin, S. A., Grisso, T., Gudjonsson, G. H., Leo, R. A., &
biomedical sciences. Behavior Research Methods, 39(2), 175–191. Redlich, A. D. (2010). Police-induced confessions: Risk factors and rec-
https://doi.org/10.3758/BF03193146 ommendations. Law and Human Behavior, 34(1), 3–38. https://doi.org/
Field, A. (2009). Discovering statistics using SPSS. Sage. 10.1007/s10979-009-9188-6
Frenda, S. J., Berkowitz, S. R., Loftus, E. F., & Fenn, K. M. (2016). Sleep Kassin, S. M., Leo, R. A., Meissner, C. A., Richman, K. D., Colwell, L. H.,
deprivation and false confessions. Proceedings of the National Academy Leach, A. M., & La Fon, D. (2007). Police interviewing and interrogation:
of Sciences of the United States of America, 113(8), 2047–2050. https:// A self-report survey of police practices and beliefs. Law and Human
doi.org/10.1073/pnas.1521518113 Behavior, 31(4), 381–400. https://doi.org/10.1007/s10979-006-9073-5
Kassin, S. M., & McNall, K. (1991). Police interrogations and confessions: Marion, S., Kaplan, J., & Cutler, B. L. (2019). Expert testimony. In
Communicating promises and threats by pragmatic implication. Law N. Brewer & A. Douglass (Eds.), Psychological science and the law
and Human Behavior, 15(3), 233–251. https://doi.org/10.1007/ (pp. 384–401). Guilford Press.
BF01061711 McAuliff, B. D., & Arter, J. L. (2016). Adversarial allegiance: The devil is
Kassin, S. M., Redlich, A. D., Alceste, F., & Luke, T. J. (2018). On the in the evidence details, not just on the witness stand. Law and Human
general acceptance of confessions research: Opinions of the scientific Behavior, 40(5), 524–535. https://doi.org/10.1037/lhb0000198
community. American Psychologist, 73(1), 63–80. https://doi.org/10 Meissner, C. A., Redlich, A. D., Bhatt, S., & Brandon, S. (2012). Interview
.1037/amp0000141 and interrogation methods and their effects on true and false confes-
Kelly, C. E., Miller, J. C., & Redlich, A. D. (2016). The dynamic nature of sions. Campbell Systematic Reviews, 8(1), 1–52. https://doi.org/10.4073/
interrogation. Law and Human Behavior, 40(3), 295–309. https://doi csr.2012.13
.org/10.1037/lhb0000172 Meissner, C. A., Redlich, A. D., Michael, S. W., Evans, J. R., Camilletti,
Kelly, C. E., Miller, J. C., Redlich, A. D., & Kleinman, S. M. (2013). A C. R., Bhatt, S., & Brandon, S. (2014). Accusatorial and information-
taxonomy of interrogation methods. Psychology, Public Policy, and gathering interrogation methods and their effects on true and false con-
Law, 19(2), 165–178. https://doi.org/10.1037/a0030310 fessions: A meta-analytic review. Journal of Experimental Criminology,
Kelly, C. E., Redlich, A. D., & Miller, J. C. (2015). Examining the meso-
10(4), 459–486. https://doi.org/10.1007/s11292-014-9207-6

level domains of the interrogation taxonomy. Psychology, Public Policy, Melzer, N. (2018). Seventieth anniversary of the Universal Declaration of
and Law, 21(2), 179–191. https://doi.org/10.1037/law0000034 Human Rights: Reaffirming and strengthening the prohibition of torture
King, L., & Snook, B. (2009). Peering inside a Canadian interrogation room: and other cruel, inhuman or degrading treatment or punishment. Interim
An examination of the Reid model of interrogation, influence tactics, and report of the Special Rapporteur on torture and other cruel, inhuman or
coercive strategies. Criminal Justice and Behavior, 36(7), 674–694. degrading treatment of punishment. https://digitallibrary.un.org/record/
https://doi.org/10.1177/0093854809335142 1639163?ln=en
King (v. Warickshall 168 Eng. Rep. 234, 235 K. B. 1783). Miranda v. Arizona 384 U.S. 436 (1966).
Klaver, J. R., Lee, Z., & Rose, V. G. (2008). Effects of personality, inter- Monahan, J. (1981). Predicting violent behavior: An assessment of clinical
rogation techniques and plausibility in an experimental false confession techniques. Sage.
paradigm. Legal and Criminological Psychology, 13(1), 71–88. https:// Munsterberg, H. (1908). On the witness stand: Essays on psychology and
doi.org/10.1348/135532507X193051 crime. Doubleday.
Knowles, E. S., & Linn, J. A. (Eds.). (2004). Resistance and persuasion.
Muraven, M., & Baumeister, R. F. (2000). Self-regulation and depletion of
Psychology Press. https://doi.org/10.4324/9781410609816
limited resources: Does self-control resemble a muscle? Psychological
Knowles, E. S., & Riner, D. D. (2007). Omega approaches to persuasion:
Bulletin, 126(2), 247–259. https://doi.org/10.1037/0033-2909.126.2.247
Overcoming resistance. In A. Pratkanis (Ed.), The science of social
Murrie, D. C., & Boccaccini, M. T. (2015). Adversarial allegiance among
influence: Advances and future progress (pp. 83–114). Psychology
expert witnesses. Annual Review of Law and Social Science, 11(1), 37–
Press.
55. https://doi.org/10.1146/annurev-lawsocsci-120814-121714
Krauss, D., & Lieberman, J. (2007). Expert testimony on risk and future danger-
Murrie, D. C., Boccaccini, M. T., Guarnera, L. A., & Rufino, K. A. (2013).
ousness. In M. Costanzo, D. Krauss, & K. Pezdek (Eds.), Expert psychologi-
Are forensic experts biased by the side that retained them? Psychological
cal testimony for the courts (pp. 227–249). Lawrence Erlbaum Associates
Science, 24(10), 1889–1897. https://doi.org/10.1177/0956797613481812
Publishers.
Murrie, D. C., Boccaccini, M. T., Turner, D. B., Meeks, M., Woods, C., &
Krauss, D. A., & Sales, B. D. (2001). The effects of clinical and scientific
Tussey, C. (2009). Rater (dis)agreement on risk assessment measures in
expert testimony on juror decision making in capital sentencing. Psy-
sexually violent predator proceedings: Evidence of adversarial alle-
chology, Public Policy, and Law, 7(2), 267–310. https://doi.org/10
giance in forensic evaluation? Psychology, Public Policy, and Law,
.1037/1076-8971.7.2.267
15(1), 19–53. https://doi.org/10.1037/a0014897
Leo, R. A. (1996). Inside the interrogation room. The Journal of Criminal
Neal, T., & Grisso, T. (2014). The cognitive underpinnings of bias in for-
Law & Criminology, 86(2), 266–303. https://doi.org/10.2307/1144028
ensic mental health evaluations. Psychology, Public Policy, and Law,
Leo, R. A. (2008). Police interrogation and American justice. Harvard
20(2), 200–211. https://doi.org/10.1037/a0035824
University Press. https://doi.org/10.4159/9780674033702
Leo, R. A., & Liu, B. (2009). What do potential jurors know about police Norvoll, R., Hem, M. H., & Pedersen, R. (2017, March). The role of ethics
interrogation techniques and false confessions? Behavioral Sciences & in reducing and improving the quality of coercion in mental health care.
the Law, 27(3), 381–399. https://doi.org/10.1002/bsl.872 HEC Forum, 29(1), 59–74. https://doi.org/10.1007/s10730-016-9312-1
Leo, R. A., & Ofshe, R. J. (1998). The consequences of false confessions: Nozick, R. (1969). Coercion. In S. Morgenbesser, P. Suppes, & M. White
Deprivations of liberty and miscarriages of justice in the age of psycho- (Eds.), Philosophy, science, and method: Essays in honor of Ernest
logical interrogation. The Journal of Criminal Law & Criminology, Nagel (pp. 440–472). St. Martin’s Press.
88(2), 429–496. https://doi.org/10.2307/1144288 Ofshe, R. J., & Leo, R. A. (1997). The social psychology of police inter-
Lidz, C. W., Mulvey, E. P., Hoge, S. K., Kirsch, B. L., Monahan, J., rogation: The theory and classification of true and false confessions.
Bennett, N. S., Eisenberg, M., Gardner, W., & Roth, L. H. (1997). The Studies in Law, Politics, and Society, 16, 189–254.
validity of mental patients’ accounts of coercion-related behaviors in the Pearse, J., & Gudjonsson, G. H. (1999). Measuring influential police inter-
hospital admission process. Law and Human Behavior, 21(4), 361–376. viewing tactics: A factor analytic approach. Legal and Criminological
https://doi.org/10.1023/A:1024803102017 Psychology, 4(2), 221–238. https://doi.org/10.1348/135532599167860
Luna, S. (2022). Defining coercion: An application in interrogation and Pepson, M. D., & Sharifi, J. N. (2010). Lego v. Twomey: The improbable
plea negotiation contexts. Psychology, Public Policy, and Law, 28(2), relationship between an obscure Supreme Court decision and wrongful
240–254. https://doi.org/10.1037/law0000345 convictions. The American Criminal Law Review, 47, 1185–1250.
Malinski v. New York 324 U.S. 401, 404 (1945). Philips, M. (1984). Are coerced agreements involuntary? Law and Philoso-
Malloy, L. C., Shulman, E. P., & Cauffman, E. (2014). Interrogations, con- phy, 3(1), 133–145. https://doi.org/10.1007/BF00211227
fessions, and guilty pleas among serious adolescent offenders. Law and R v. Oickle 2 SCR 3, SCR 38 (2000).
Human Behavior, 38(2), 181–193. https://doi.org/10.1037/lhb0000065 R v. Piche SCR 23 (1971).
Redlich, A. D. (2004). Law & psychiatry: Mental illness, police interroga- Thomas, G. C., III, & Leo, R. A. (2014). Law of police interrogation. In G.
tions, and the potential for false confession. Psychiatric Services, 55(1), Bruinsma & D. Weisburd (Eds.), Encyclopedia of criminology & crimi-
19–21. https://doi.org/10.1176/appi.ps.55.1.19 nal justice (pp. 2835–2840). Springer.
Redlich, A. D. (2007). Double jeopardy in the interrogation room for Tickle-Degnen, L., & Rosenthal, R. (1990). The nature of rapport and its
youths with mental illness. American Psychologist, 62(6), 609–611. nonverbal correlates. Psychological Inquiry, 1(4), 285–293. https://doi
https://doi.org/10.1037/0003-066X62.6.609 .org/10.1207/s15327965pli0104_1
Redlich, A. D. (2009). The susceptibility of juveniles to false confessions Trainum, J. L. (2016). How the police generate false confessions: An
and false guilty pleas. Rutgers Law Review, 62, 943–957. inside look at the interrogation room. Rowman & Littlefield.
Redlich, A. D., Kulish, R., & Steadman, H. J. (2011). Comparing true and Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging
false confessions among persons with serious mental illness. Psychol- frequency and probability. Cognitive Psychology, 5(2), 207–232. https://
ogy, Public Policy, and Law, 17(3), 394–419. https://doi.org/10.1037/ doi.org/10.1016/0010-0285(73)90033-9
a0022918 Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuris-
Reid Institute. (2015). International research validates the core elements tics and biases. Science, 185(4157), 1124–1131. https://doi.org/10.1126/
of the Reid technique. http://www.reid.com/pdfs/20150511.pdf science.185.4157.1124
Rufino, K. A., Boccaccini, M. T., Hawes, S. W., & Murrie, D. C. (2012). Tversky, A., & Kahneman, D. (1986). The framing of decisions and the
evaluation of prospects. In R. Barcan Marcus, G. J. W. Dorn, &

When experts disagreed, who was correct? A comparison of PCL-R
P. Weingartner (Eds.), Studies in logic and the foundations of math-

scores from independent raters and opposing forensic experts. Law and
ematics (pp. 503–520). Elsevier.
Human Behavior, 36(6), 527–537. https://doi.org/10.1037/h0093988
U.S. Const., amend. II.
Schatz, S. J. (2018). Interrogated with intellectual disabilities: The risks of
Vallano, J. P., & Schreiber Compo, N. (2015). Rapport-building with co-
false confession. Stanford Law Review, 70(2), 643–690.
operative witnesses and criminal suspects: A theoretical and empirical
Scherr, K. C., Redlich, A. D., & Kassin, S. M. (2020). Cumulative disad-
review. Psychology, Public Policy, and Law, 21(1), 85–99. https://doi
vantage: A psychological framework for understanding how innocence
.org/10.1037/law0000035
can lead to confession, wrongful conviction, and beyond. Perspectives
Vrij, A., Fisher, R., Mann, S., & Leal, S. (2006). Detecting deception by
on Psychological Science, 15(2), 353–383. https://doi.org/10.1177/ manipulating cognitive load. Trends in Cognitive Sciences, 10(4), 141–
1745691619896608 142. https://doi.org/10.1016/j.tics.2006.02.003
Schulhofer, S. J. (2001). Miranda, Dickerson, and the puzzling persistence Wachi, T., & Lamb, M. E. (2019). Japanese public opinion about suspect
of Fifth Amendment exceptionalism. Michigan Law Review, 99(5), interviewing techniques. Legal and Criminological Psychology, 24(1),
941–957. https://doi.org/10.2307/1290522 55–70. https://doi.org/10.1111/lcrp.12139
Smith, S., Stinson, V., & Patry, M. (2012). Confession evidence in Canada: Watts, S. E., & Weems, C. F. (2006). Associations among selective atten-
Psychological issues and legal landscapes. Psychology, Crime & Law, tion, memory bias, cognitive errors and symptoms of anxiety in youth.
18(3), 317–333. https://doi.org/10.1080/1068316X.2010.486380 Journal of Abnormal Child Psychology, 34(6), 841–852. https://doi.org/
Soree, N. (2005). When the innocent speak: False confessions, constitu- 10.1007/s10802-006-9066-3
tional safeguards, and the role of expert testimony. American Journal of Wickersham Commission. (1931). Report on lawlessness in law enforcement.
Criminal Law, 32, 191–263. U.S. Government Printing Office.
Starr, D. (2019, June 13). This psychologist explains why people confess to Wicklander-Zulawski & Associates. (2020). Interviewing techniques.
crimes they didn’t commit. Science Magazine. https://www.sciencemag
.org/news/2019/06/psychologist-explains-why-people-confess-crimes-
they-didn-t-commit Received March 23, 2022
Thomas, G. C., III, & Leo, R. A. (2009). Confessions of guilt: From tor- Revision received July 10, 2022
ture to Miranda and beyond. Oxford University Press. Accepted July 11, 2022 n

Development of The Interview and Interrogation Assessment Instrument

Uploaded by

Copyright:

Available Formats

Development of The Interview and Interrogation Assessment Instrument

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Development of The Interview and Interrogation Assessment Instrument

Uploaded by

Copyright:

Available Formats

Psychology, Public Policy, and Law

© 2022 American Psychological Association 2023, Vol. 29, No. 1, 46–61

Development of the Interview and Interrogation Assessment Instrument

Jeffrey Kaplan1, Brian L. Cutler2, Amy-May Leach1, and Joseph Eastwood1

unstructured professional judgment. To address this lack of standardization, we have developed a

Keywords: interrogations, false confessions, expert witness, coercion

Important factors may be overlooked and/or inﬂuenced by bias. In

certain maximization techniques (Kassin & McNall, 1991) that

to aid those who already possess a mastery of the psychology of

Luna’s (2022) deﬁnition. The ability to resist the demands of

ing the domains, the general structure of meso-level domains that

they identiﬁed has been retained in the IIAI.

and by the ratings given by experts. Based on those verbal and

to the point that it could be considered good to excellent. Summar-

Table 4 13.9% (n = 5) possession of child pornography and/or soliciting

study. Those who were interested in participating were invited to

click on a Qualtrics link, which forwarded them to the letter of

or emphasizing the certainty of guilt or conviction; and (e) made Discussion

appeals to the suspect’s self-interest, offered excuses or justiﬁca-

10(4), 459–486. https://doi.org/10.1007/s11292-014-9207-6

evaluation of prospects. In R. Barcan Marcus, G. J. W. Dorn, &

P. Weingartner (Eds.), Studies in logic and the foundations of math-

You might also like