Use of The Psychopathy Checklist-Revised in Legal Contexts Validity, Reliability

Journal of Personality Assessment
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/hjpa20
Use of the Psychopathy Checklist-Revised in Legal

Contexts: Validity, Reliability, Admissibility, and
Evidentiary Issues
David DeMatteo & Mark E. Olver
To cite this article: David DeMatteo & Mark E. Olver (2022) Use of the Psychopathy Checklist-
Revised in Legal Contexts: Validity, Reliability, Admissibility, and Evidentiary Issues, Journal of
Personality Assessment, 104:2, 234-251, DOI: 10.1080/00223891.2021.1955693
To link to this article: https://doi.org/10.1080/00223891.2021.1955693
Published online: 06 Aug 2021.
Submit your article to this journal
Article views: 4003
View related articles
View Crossmark data
Citing articles: 2 View citing articles
Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=hjpa20
JOURNAL OF PERSONALITY ASSESSMENT
2022, VOL. 104, NO. 2, 234–251
https://doi.org/10.1080/00223891.2021.1955693
Use of the Psychopathy Checklist-Revised in Legal Contexts: Validity, Reliability,

Admissibility, and Evidentiary Issues
David DeMatteo1 and Mark E. Olver2
1
Department of Psychology & Thomas R. Kline School of Law, Drexel University, Philadelphia, Pennsylvania; 2Department of Psychology and
Health Studies, University of Saskatchewan, Saskatoon, Saskatchewan, Canada
ABSTRACT ARTICLE HISTORY

The construct of psychopathy has received considerable attention from clinicians, researchers, and Received 16 April 2021
legal practitioners because of its demonstrated association with a range of outcomes of interest to Accepted 29 June 2021
the criminal justice system. The Psychopathy Checklist-Revised (PCL-R) is generally regarded as the
premier assessment tool for measuring psychopathy in correctional and legal contexts, and the
PCL-R is being used with increased frequency to address a variety of legal questions. This article
provides a comprehensive examination and review of the PCL-R’s use in legal contexts. We begin
by reviewing various uses (appropriate and inappropriate) of the PCL-R in legal contexts, using the
risk-need-responsivity (RNR) model as the conceptual framework. After reviewing available data
regarding the use of the PCL-R in legal contexts, we review and synthesize psychometric research
with psycholegal relevance, with a focus on the PCL-R’s construct validity, predictive validity, and
interrater reliability. We then discuss the scientific acceptability and clinical utility of the PCL-R’s
structural, predictive, and measurement properties for credibility in court, followed by sample
cross-examination questions. We conclude with a review of admissibility issues relating to the use
of the PCL-R in various legal proceedings.
Psychopathy is a serious personality disorder with criminal detailed clinical interview (using a semi-structured interview
justice relevance. Representing a constellation of problematic guide) and review of archival (e.g., institutional files) and col-
interpersonal (e.g., superficiality, deceitfulness, manipulation), lateral information. Factor analytic research indicates that the
affective (e.g., lack of remorse, callous, shallow affect), and PCL-R is composed of two factors, with Factor 1 consisting
behavioral features (e.g., impulsivity, irresponsibility, poor of interpersonal/affective features and Factor 2 consisting of
behavior controls, criminal versatility), psychopathy is associ- behavioral features (e.g., Harpur et al., 1989), and the two
ated with a miscellany of adverse criminal justice outcomes PCL–R factors have been further subdivided into facets: inter-
(Guy et al., 2005; Hare, 2016; Leistico et al., 2008) and people personal, affective, lifestyle, and antisocial (Hare, 2003).
with large numbers of these features tend to This article examines and reviews the use of the PCL-R
be overrepresented in correctional, forensic, and justice system in legal contexts. We begin by reviewing applications of
contexts. Psychopathy is routinely assessed in legal contexts— the PCL-R in legal settings conceptualized around the risk-
most frequently via the Psychopathy Checklist-Revised (PCL- need-responsivity (RNR) model, with illustrations of com-
R; Hare, 1991, 2003)—and used to inform decision making by mon appropriate and inappropriate uses, followed by a
the trier of fact (DeMatteo et al., 2014b, 2016, 2019). review of clinician surveys and case law reviews of PCL-R
The PCL-R is a 20-item symptom construct rating scale use in legal contexts. We then review and synthesize psycho-
for the assessment of psychopathy. With items rated on a metric research with psycholegal relevance, including the
three-point ordinal scale (0 not present, 1 possibly/partially PCL-R’s construct validity as a measure of psychopathic per-
present, 2 present), total scores range from 0-40 and represent sonality, its predictive validity for community and institu-
the extent to which the individual represents a prototypically tional recidivism, and its interrater reliability in controlled
psychopathic individual. Although psychopathy is a dimen- research settings and field contexts. Putting these psycho-
sional construct measured by a dimensional tool (Edens et al., metric research findings into context, we comment on the
2006), cut scores of 30 have traditionally been used to charac- threshold of scientific acceptability and clinical utility of the
terize psychopathy, although lower cut scores (e.g., 25 or 26) PCL-R’s structural, predictive, and measurement properties
have been used in research contexts via file review (e.g., for credibility in court. We follow with 10 key questions
Wong, 1988). The PCL-R is typically rated based on a for cross examination on the PCL-R. We conclude with a
CONTACT David DeMatteo [email protected] Department of Psychology, Drexel University, 3141 Chestnut St., Stratton Suite 119, Philadelphia,
PA 19104, USA; Mark E. Olver [email protected] University of Saskatchewan, Department of Psychology and Health Studies, 9 Campus Drive, Arts
Building Room 154, Saskatoon, SK, S7N 5A5, Canada.
This is a shared lead authorship article with the contributors listed in alphabetical order.
ß 2021 Taylor & Francis Group, LLC
USE OF THE PSYCHOPATHY CHECKLIST-REVISED IN LEGAL CONTEXTS 235
review of admissibility of the PCL-R in court, including a evaluate changes in risk from treatment or other change
discussion of instances in which the threshold of evidence agents to aid release decisions and case planning.
may support (or not support) the admissibility of the PCL-R Finally, per the responsivity principle (i.e., individualize
as evidence in certain types of legal proceedings. service delivery to unique client characteristics), high PCL-R
scorers are likely to pose challenges to service providers in
motivation, engagement, cooperation, interpersonal dynam-
Applications of the PCL-R in legal settings ics and so forth that need to be managed and monitored as
A risk-need-responsivity conceptualization of PCL-R part of routine services. Factor 1, in particular, embodies a
measured psychopathy set of responsivity issues; readers will be familiar with the
“litigious patient” and treatment interfering dynamics such
The PCL-R can be viewed as tool to inform risk-need- as dominating or derailing group, manipulating or intimi-
responsivity (RNR) applications (per Andrews & Bonta, dating staff or co-patients, and splitting staff. These issues,
1994–2010; Bonta & Andrews, 2017) that performs well clin- however, can be readily managed and are essentially a reflec-
ically and psychometrically in measuring a forensically rele- tion of the clinical issues that got the individual referred for
vant psychopathology construct when: 1) raters are services in the first place (why would we expect any differ-
adequately trained, 2) have access to quality information, ently?). The wisdom of Wong’s (2016) 2-Component (2-C)
and 3) are not laboring under systemic pressures or biases
model is to appropriately expect, tolerate, and manage
that can compromise accuracy of ratings (e.g., adversarial
Factor 1 issues (which will usually improve), per the respon-
allegiance; see Boccaccini et al., 2008, 2012). The RNR
sivity principle, while directly treating Factor 2 issues, per
model is a dominant paradigm for the integration of correc-
the risk and need principles.
tional/forensic assessment and treatment to reduce and
manage risk, and PCL-R measured psychopathy can be read-
ily incorporated into this framework. Legal applications
First, per the risk principle (i.e., service intensity matched
to risk level), high PCL-R scorers are likely candidates for The PCL-R is widely used for a range of legal applications
high intensity services (they should be top priority) given that can be integrated within the RNR framework, some of
the PCL-R’s association with recidivism (see “Predictive which remain the source of debate among experts in the
Validity” section), but the PCL-R should not be used to field. We review a few key illustrative examples of how the
determine risk level or treatment intensity in isolation. PCL-R is commonly used.
Although the PCL-R is not a risk assessment tool, it is fre- Presentence evaluations. The PCL-R may be used in a
quently used for this purpose given its predictive properties. pretrial or presentence context to inform decisions such as
For instance, an individual with a high PCL-R score, bail versus remand, community restrictions, or sentencing.
coupled with Average-Medium risk on specialized sexual In this context, the PCL-R is most frequently applied as a
violence measures such as the Static-99R (Helmus et al., risk measure, in which high scores would indicate greater
2012) and Sexual Violence Risk-20 (SVR-20; Boer et al., concern about the potential for pretrial or presentence recid-
1997), may require a moderate intensity sexual offense treat- ivism (e.g., witness intimidation, breaches, commission of
ment program. The same person, however, may score quite new crimes), and may be used as one piece of evidence to
high on a general risk-needs tool such as the Level of inform remand decisions; or alternatively, raise similar con-
Service/Case Management Inventory (LS/CMI; Andrews cerns about the potential for post-sentence recidivism. In
et al., 2004), and still require high intensity community either instance, the PCL-R should be paired with another
supervision upon release to prevent technical violations and formalized measure of risk and need (integrated with other
promote smooth reintegration, even if sexual violence risk is information sources) in the context of a comprehensive
lower and quite manageable. assessment to guide such decision making; another measure
Second, per the need principle (i.e., prioritize crimino- will help contextualize the risk and give some indication
genic needs for risk reducing intervention), high PCL-R about how to manage the risk (e.g., areas for monitoring or
scorers have a greater density and severity of criminogenic to advise community restrictions while on bail, or for inter-
needs, which require a comprehensive suite of services pri- vention post sentence).
oritized around their areas of risk and need (Polaschek & Preventative detention. The PCL-R is frequently used to
Wong, 2020). For instance, PCL-R scores, especially Factor inform decisions regarding applications for preventative
2, correlate highly with formalized dynamic risk measures detention such as Dangerous Offender (DO, in Canada) and
such as the Violence Risk Scale (VRS; Wong & Gordon, Sexual Offender Civil Commitment (SOCC, in certain U.S.
1999–2003), Historical Clinical Risk-20 (HCR-20; Douglas jurisdictions) evaluations, or capital sentencing (i.e., death
et al., 2013), and measures of the Central Eight need penalty). The use of the PCL-R in the context of a DO
domains (i.e., antisocial history, attitudes, associates, and evaluation can inform decisions (taken in consideration with
personality pattern, employment/education, family marital, the larger body of legal evidence) such as provision of a
substance abuse, and leisure/recreation) (e.g., Douglas et al., determinate or indeterminate sentence, or alternatively,
2005; Simourd & Hoge, 2000; Wong & Gordon, 2006). A another form of preventative detention such as a long-term
dynamic risk measure (e.g., LS/CMI, HCR-20, VRS), how- offender (LTO) or long-term supervision order (LTSO). In
ever, is still required to identify where to intervene and to the context of SOCC evaluations, the PCL-R is frequently
236 DEMATTEO AND OLVER
used as part of an assessment conducted at the point of sen- the PCL-R are appropriate to the extent that they are con-
tence expiry to evaluate the potential for sexual violence and sistent with an RNR conceptualization of PCL-R measured
existence of a mental disorder linked to sexual violence (see psychopathy; however, when applications veer from this
DeMatteo et al., 2014a). The resulting assessment can framework and into territory that is unchartered scientific-
inform civil commitment decisions as a Sexually Violent ally or clinically, they become increasingly inappropriate.
Predator or Person (SVP; fully committed or probable Appropriate use. Broadly speaking, appropriate uses of
cause), and a resulting period of hospitalization, with release the PCL-R involve applications within the RNR model to
contingent on demonstrated reductions in risk or successful promote risk management and staff, inmate/patient, and
legal appeal. community safety and wellbeing. First, the PCL-R should be
Finally, the PCL-R has been used in capital sentencing used as part of a comprehensive clinical evaluation along
evaluations with much debate. One author (DD) of this with at least one risk measure that contains dynamic risk
article concludes that it is unwise and perilous to use the factors (e.g., HCR-20, VRS, LS/CMI) to corroborate (or dis-
PCL-R in a capital sentencing evaluation (i.e., validity and confirm) the level of risk and to identify targets for risk
reliability literature not sufficiently compelling, potential for management. The risk measure should be specific to the
misuse too high) (see DeMatteo et al., 2020), while the other type of outcome risk being evaluated (e.g., sexual offending,
author (MEO) suggests the instrument could still have its general violence, intimate partner violence [IPV], general
place if responsibly used (i.e., in tandem with other meas- reoffending).
ures to guide risk management) (see Olver et al., 2020). Second, the PCL-R can be used with appraisals of respon-
Both authors, however, firmly believe that the PCL-R should sivity considerations and other relevant clinical considera-
not be used to make life or death recommendations about a tions, such as measures of cognitive/intellectual functioning,
defendant or to provide definitive statements about an indi- academic achievement, personality and psychopathology
vidual’s propensity for violence in custody; to do so would (e.g., MMPI measures), or formalized measures of strengths/
constitute an inappropriate and arguably unethical use of protective factors (e.g., Structured Assessment of Protective
the PCL-R. Factors (SAPROF); de Vogel et al., 2009). This can help
Correctional Intake and Case Management. The PCL-R contextualize the PCL-R score and concordant sentencing,
can also be used as part of correctional intake assessments case planning, and risk management considerations; for
and ongoing correctional case management. Relevant exam- instance, a PCL-R score (whether it be high, low, or some-
ples include security classification (e.g., institutional trans- where in the middle) can take on a different meaning and
fers, security level changes) and correctional program impact decision making and service delivery if the individ-
planning, particularly service intensity (e.g., low vs. moder- ual, for example, has schizophrenia, a neurodevelopmental
ate vs. high intensity) and at a broad level, the breadth and disorder (e.g., FASD), and/or a number of strengths and
foci of services (i.e., a high PCL-R score may bode for com- resiliencies (e.g., intellectual ability, prosocial supports, posi-
prehensive services, while a low PCL-R score may require tive coping). Further, Olver and Riemer (2021) found men
lower intensity services) per the risk and need principles. As with high PCL-R scores who also had several protective fac-
noted previously, the PCL-R can also be used to identify tors on the SAPROF had lower rates of violent recidivism
special responsivity considerations for service providers to than high PCL-R scoring men with few protective factors.
address in correctional programming (e.g., managing treat- Third, the totality of the PCL-R profile—including factors
ment interfering behaviors to promote client retention and and facets—can inform services and decision making. For
treatment gain). instance, given the association between Factor 1 and treat-
Prerelease. Finally, the PCL-R is frequently used to ment behavior (see Wong, 2016, for a review), per the
inform release decision making, such as applications for responsivity principle, prominent interpersonal and affective
parole or other forms of conditional, supervised release features on the PCL-R means that special efforts will likely
(DeMatteo et al., 2014b). This includes evaluations of release be needed to engage the person in treatment (e.g., motiv-
suitability and the extent to which the individual can be ational interviewing, focusing on task and goal components
safely managed in the community. PCL-R scores have robust of the working alliance, monitoring countertransference,
support as a valid predictor of recidivism among those maintaining boundaries, teamwork).
released from incarceration (e.g., Leistico et al., 2008), and Fourth, we encourage use of percentile ranks and the
an item analysis of the PCL-R, particularly when combined appropriate normative group to contextualize PCL-R total,
with other tools that can identify criminogenic needs that factor, and facet scores, in addition to, or in replacement of,
can be targeted for intervention, can provide important risk- actual numeric score(s). For instance, a PCL-R total score
and treatment-relevant information that can inform release two SDs below the mean (i.e., around 6, or the 2nd percent-
decision making (see Loving, 2002). ile) for an adult male prison inmate is equally informative
as one that is close to two SDs above the mean (i.e., around
37 or 38, or the 98th percentile); the case management
Best practices for legal applications
implications, however, should be very different.
There are several considerations for appropriate and Fifth, and finally, all the above is taken together with pro-
inappropriate use of the PCL-R in the aforementioned legal moting competent and ethical use of the PCL-R, including
contexts to inform best practices. In our view, the uses of ensuring that the evaluator has been adequately trained by a
qualified trainer, has access to quality, detailed information primary limitations. First, they are focused exclusively on
for completing PCL-R ratings and a proper assessment, and PCL-R use in court cases, which means that use of the PCL-
is alert to potential adversarial biases or allegiance effects R in many clinical or treatment contexts will not be cap-
that could impact the scoring and interpretation of the PCL- tured. Second, because commonly used legal databases such
R and other measures. Evaluators should try to refrain from as Westlaw and Lexis-Nexis typically only contain court
use of the label “psychopath,” replaced with person-first lan- opinions that have reached the appellate stage of litigation,
guage when employing the tool in these contexts (e.g., a per- the results of case law reviews underestimate PCL-R use in
son with substantial psychopathic features). court because most cases involving the PCL-R in U.S. courts
Inappropriate uses. There are a number of positive and are state-level cases that do not result in a written opinion
helpful uses of the PCL-R, but there are also several ways to included in electronic legal databases. Third, some states
inappropriately use the measure, such as to label a person a and provincial systems mandate use of the PCL-R for cer-
“psychopath,” a damaging and stigmatizing moniker that is tain legal issues (e.g., parole decisions, SVP determinations),
pejorative in its connotations, potentially prejudicial in its and these uses of the PCL-R are difficult to track and not
impact, and contrary to the wisdom of general responsivity adequately captured by case law reviews.
and core correctional practice. Also inappropriate would be Singh et al. (2014) conducted the largest and most com-
to use the PCL-R as a standalone tool to evaluate risk or dan- prehensive international clinician survey that included data
gerousness, to use it only with a static measure, or to argue on PCL-R use. In their study, they examined violence risk
on the basis of a PCL-R score that an individual will inevit- assessment practices among 2,135 practitioners who regu-
ably reoffend violently or in any other category. As the review larly conduct violence risk assessments from 44 countries.
of psychometric research on the PCL-R will demonstrate, the Among all survey respondents, the most frequently used
tool predicts some outcomes considerably better than others, measure of violence risk assessment, both lifetime and in the
and as with all tools, there are limits to its predictive validity 12 months immediately preceding the survey, was the HCR-
as well as factors that can undermine reliability. Further, lim- 20 (Douglas et al., 2013), followed by the PCL-R. With
ited use of the PCL-R, especially with a static tool only, respect to lifetime use among respondents from the United
results in a narrow, singularly focused assessment that pro- States (13% of the survey respondents), the PCL-R was the
vides an incomplete picture of risk, overlooks important treat- most frequently used measure, followed by the HCR-20.
ment needs, and cannot capture change. Finally, with respect to use of violence risk assessment meas-
A further, yet common, inappropriate practice has been ures in the 12 months prior to the survey, the most fre-
to use the PCL-R to argue that someone is untreatable, to quently used measure among respondents from the United
deny an individual access to services, or alternatively, using States was the HCR-20, followed by the LSI-R (Andrews &
the tool to generate an “attrition profile” to weed the indi- Bonta, 1995), Violence Risk Appraisal Guide (VRAG;
vidual out of treatment, contrary to the risk and need prin- Quinsey et al., 2006), and PCL-R.
ciples. As we will review further in this article, there is an There have been several case law reviews examining use
encouraging and increasingly robust literature to show that
of the PCL-R in legal proceedings in the United States, the
high psychopathy individuals can respond to RNR-based
United Kingdom, and Canada (DeMatteo et al., 2014b;
treatment and risk management services (Polaschek &
DeMatteo & Edens, 2006; Gagnon et al., 2007; Howard
Wong, 2020; Sewall & Olver, 2019; Skeem et al., 2002;
et al., 2012; Lloyd et al., 2010; Viljoen et al., 2010). The first
Wong et al., 2012). Finally, as already noted but worth
PCL-R systematic case law reviews were published in 2006
repeating, it is inappropriate to use the PCL-R to make life
and included United States court cases published between
or death recommendations about an individual for ethical,
January 1, 1991 (commercial publication date of the PCL-R)
humanitarian, and psychometric reasons. We return to this
and December 31, 2004 (DeMatteo & Edens, 2006; Walsh &
issue toward the end of the article in the broader context of
Walsh, 2006). DeMatteo and Edens (2006) identified 87
institutional risk management considerations for persons
published cases (76 state; 11 federal) involving the PCL-R,
with high PCL-R scores.
and Walsh and Walsh (2006) identified 76 cases (67 state; 9
federal) involving the PCL-R. A more recent case law review
Use of the PCL-R revealed that the PCL-R is being used with increasing fre-
quency in United States courts (DeMatteo et al., 2014b).
Clinician surveys and case law reviews
Specifically, in a follow-up to the DeMatteo and Edens
Data on PCL-R use primarily come from clinician surveys (2006) study, DeMatteo et al. (2014b) found 348 published
and systematic reviews of court cases. There are, however, cases (304 state and 44 federal) involving the PCL-R
limitations in both sources of data that restrict our ability to between 2005 and 2011 (inclusive), which is a 300þ%
make firm conclusions regarding PCL-R use. For example, increase in a time period half as long as the time period in
although surveys provide useful data from practicing clini- the prior study. In all of these case law reviews, the PCL-R
cians in a variety of settings, they are limited by several fac- was primarily used as a risk assessment tool introduced by
tors, including sampling considerations (e.g., sample size, the prosecution.
recruitment method), the inability to determine a response Several researchers have examined use of the PCL-R (or
rate, and respondent errors (e.g., misunderstanding of a sur- derivative measure) in Canadian cases. Viljoen et al. (2010)
vey item, errors in data entry). Case law reviews have three examined the use of psychopathy measures (including the
PCL) in United States and Canadian legal proceedings specifically convergent, discriminant, and factorial validity;
involving adolescent offenders, and they identified 111 cases 2) Predictive (criterion-related) validity of PCL-R test scores
(71 Canadian; 40 U.S.) involving 143 separate evaluations for criminal justice criteria that are of legal relevance (e.g.,
involving an assessment of psychopathy. In their review of future violence, institutional misconduct); and 3) Interrater
written decisions or oral transcripts from 136 dangerous reliability of independent PCL-R ratings completed in
offender (or long-term offender) hearings in Canada to laboratory and field settings.
examine whether use of the PCL-R was related to trial out-
come, Lloyd et al. (2010) found that the PCL-R was used in
58 (42.6%) of the cases. Construct validity
Psychometric research examining the construct validity of
Admissibility challenges to the PCL-R the PCL-R has demonstrated support for the stability of its
factor structure as well as theoretically and clinically mean-
There are limited data, drawn from case law reviews and ingful convergent and discriminant associations with opera-
practitioner surveys, regarding how often the introduction tionalizations of pertinent constructs. Since the inaugural
of the PCL-R in court is challenged. In all United States two-factor solution was identified on the original 22-item
jurisdictions, the admissibility of evidence is governed by PCL, research has affirmed the two factor and subsequently
rules of evidence, and the evidentiary rules in all jurisdic- a correlated four-facet model (employing 18 of the 20 PCL-
tions require that evidence be both relevant and non-preju- R items) comprising interpersonal (e.g., superficial charm,
dicial/non-misleading as a threshold for admissibility (see grandiose, deceitful) and affective (e.g., lack of remorse, cal-
Federal Rules of Evidence, 1975). (See section titled lous, shallow affect) facets (subsumed by Factor 1), and life-
“Admissibility of the PCL-R” for a detailed discussion of style (e.g., parasitic, impulsivity, irresponsibility) and
admissibility standards.) The available data, albeit quite lim- antisocial (e.g., poor behavioral controls, early behavior
ited, suggest that the admissibility of the PCL-R is rarely problems, criminal versatility) facets (subsumed by Factor 2)
challenged and that challenges are rarely successful
(Hare, 2003). For instance, Neumann et al.’s (2015) struc-
(DeMatteo et al., 2014b; Edens & Cox, 2012).
tural equation modeling of international PCL datasets repli-
Edens and Cox (2012) reported the results of a survey of
cated the four-facet model for the PCL-R in large male
attendees (N ¼ 41) at a national capital mitigation confer-
correctional and forensic samples from North America,
ence. Most respondents (n ¼ 21) were defense attorneys,
South America, Europe, Netherlands, Austria, Sweden, and
with the remainder including mitigation specialists (n ¼ 14)
Finland, as well as a North American female sample (Ns ¼
and other professionals involved in capital cases (e.g., inves-
209 to 4,685, Relative Fits ¼ .90 to .97). A 13-item three-fac-
tigators). Approximately 60% of the respondents reported
that the admissibility of expert evidence labeling a defendant tor model excluding the antisocial factor was also identified
as a “psychopath,” “sociopath,” or “antisocial” had been (Cooke & Michie, 2001); however, this model does not
legally challenged, but that such challenges were rarely suc- examine the latent structure of all PCL-R items, so it repre-
cessful. The survey was not exclusively focused on PCL-R sents a conceptual model of psychopathy (i.e., one in which
use, but a sizeable portion of respondents indicated the antisociality is not an inherent feature) rather than the latent
PCL-R was used to assess antisocial, sociopathic, or psycho- structure of the full 20-item tool (Neumann et al., 2007).
pathic traits by prosecution expert witnesses. PCL-R total, Factor 1, and Factor 2 scores have demon-
In their case law review, DeMatteo et al. (2014b) reported strated convergent associations with clinical measures of
that legal challenges to the admissibility of the PCL-R psychopathy, such as the Comprehensive Assessment of
occurred in only 6% of the identified cases, or 21 times Psychopathic Personality (Sandvik et al., 2012), self-report
across 19 of the 348 cases they found. The defense chal- measures of psychopathology such as the Personality
lenged the admissibility of the PCL-R in 18 of those 19 Assessment Inventory (Edens et al., 2000; Gardner &
cases, most commonly on the grounds that the PCL-R had Boccaccini, 2017), Big Five personality dimensions such as
little or no relevance to the issue before the court; challenges low agreeableness and conscientiousness (Lilienfeld et al.,
based on the PCL-R’s potential to be prejudicial were very 2015), and self-report measures of psychopathy such as the
rare (two cases). They reported that admissibility challenges Self-Report Psychopathy (SRP) measures (Hollerbach et al.,
were successful in four of the 19 cases. 2020; Sandvik et al., 2012), the Psychopathic Personality
Inventory, and Levenson’s Primary and Secondary
Psychopathy measures (Poythress et al., 2010). Other evi-
Psychometric properties of PCL-R scores dence for convergent and discriminant validity comes from
The PCL-R has amassed a considerable evidence base sup- positive correlations with prototypicality ratings of anti-
porting the validity of inferences made on the basis of test social, narcissistic, and histrionic personality disorder (Hart
scores as a measure of the psychopathy construct, as well as & Hare, 1989), substance use diagnoses (Hart & Hare, 1989,
evidence for the consistency of test scores in repeated meas- Hemphill et al., 1994), and negative associations with proto-
urements (i.e., reliability) of this construct. This review typicality ratings of avoidant and dependent personality dis-
addresses three sets of psychometric properties of the PCL-R order (Hart & Hare, 1989) and (notably Factor 1)
with psycholegal relevance: 1) Construct validity, and more neuroticism (Hollerbach et al., 2020; Lilienfeld et al., 2015).
Predictive validity the youth variant, the Psychopathy Checklist: Youth Version
(PCL: YV; Forth et al., 2003), given that the focus of the pre-
Predictive validity research on the PCL-R has primarily
sent review is on the adult version of the PCL-R measures.
focused on relative risk or discrimination; that is, to what
As seen in Table 1, most meta-analyses of the PCL-R
extent members of a criterion or reference group, such as
have examined, in aggregate, predictive effects for the ori-
recidivists, can be accurately differentiated from other cases
ginal PCL (Hare, 1980), its Screening Version (PCL: SV;
not part of the reference group (i.e., non-recidivists) based on
Hart et al., 1995), and in some instances, even the Youth
their PCL-R scores. This is most frequently evaluated through
Version; the lone exception is Hawes et al. (2013), who took
receiver operator characteristic curve (ROC) analyses (area
note of this issue with previous meta-analyses and purpos-
under the curve or AUC), point biserial correlation (rpb), or
ively only examined the PCL-R and not the youth or screen-
standardized mean difference (Cohen’s d). Rice and Harris
ing versions. These issues notwithstanding, the PCL-R
(2005) provide effect size (ES) guidelines for the association
proper generally constitutes the vast majority of studies in
of test scores with binary recidivism criteria using these three
generating the reported effects. When the PCL: SV and
metrics corresponding to effects that are small/low (AUC ¼
PCL-R had separate effects, for consistency with previous
.56, r ¼ .10, d ¼ .20), medium/moderate (AUC ¼ .64, r ¼
meta-analyses, a weighted average effect was taken. Table 1
.24, d ¼ .50), and large/high (AUC ¼ .71, r ¼ .37, d ¼ .80).
Such metrics provide some consistency and help reduce arbi- then reports a meta-analytic aggregate of standardized mean
trariness when describing the strength and magnitude of pre- difference (d), the median d and mean unweighted d, for
diction effects. An important consideration when interpreting the PCL-R total, Factor 1, and Factor 2 score with respect to
these effect sizes is that the magnitudes of some (such as r) the six outcome criteria.
are attenuated by extreme base rates of the criterion (most There are several themes evident from Table 1. First,
frequently low base rates for uncommon phenomena, such as PCL-R total scores evince broadly medium effects for the
serious violence), all metrics are attenuated by lack of vari- prediction of most recidivism criteria, including general, vio-
ance in test scores (i.e., range restriction), and power will be lent, and IPV recidivism, as well as general institutional mis-
diminished by small sample sizes and extreme base rates conducts (median and mean unweighted d > .50). PCL-R
(Babchishin & Helmus, 2016). total scores evince small effects for sexual recidivism and
The PCL-R has marshaled evidence for its predictive valid- small to medium effects for institutional violence. Second,
ity for a range of criminal justice criteria, including violation Factor 2 has larger prediction effects than Factor 1 and, as a
of conditional release (Hart et al., 1988), lengthy criminal rule, had the same magnitude of prediction effects for each
careers (Olver & Wong, 2015), institutional adjustment and of its criterion variables as the PCL-R total score (i.e.,
security risk (Olver et al., 2020), and both instrumental and broadly medium for general, violent, and institutional
reactive violence (Blais et al., 2014). PCL-R measured psych- offending; small for institutional violence and sexual recidiv-
opathy also has robust associations with criminal recidivism. ism). Factor 1 evinced small effects for all outcome criteria
Over the past 25 years, several meta-analyses have examined with the exception of weak (i.e., subthreshold) effects for
the predictive validity of the PCL-R and its variants for com- sexual recidivism. Third, the median and unweighted d val-
munity and institutional recidivism outcomes, on its own or ues are either identical or close in magnitude for supporting
in tandem with other forensic risk assessment measures. the robustness and stability of these effects across multiple
Olver et al. (2020) conducted a meta-meta-analysis of seven meta-analyses.
meta-analyses of the PCL-R and its derivatives for the predic- The meta-meta-analysis admittedly focused on the pre-
tion of different institutional outcomes, including institutional dictive properties of standalone PCL-R scores, which natur-
violence, general aggression (e.g., verbal and/or physical), and ally stands in contrast to how we advise using it. A number
institutional infractions in general. They also conducted an of studies have examined the incremental predictive validity
updated meta-analysis of new studies published within the of PCL-R scores for recidivism beyond the scores of formal
past 10-15 years not included in previous PCL-R institutional risk assessment measures, with the evidence varying with
behavior meta-analyses. the tool, sample size, and type of outcome. For instance,
We extended the work of Olver et al. (2020) by summariz- research has demonstrated PCL-R scores to increment the
ing the meta-analytic effects across 18 PCL-R related meta- prediction of general community violent recidivism control-
analyses of its predictive validity for general (k meta-analyses ling for the VRS (Olver et al., 2013) and its sexual violence
or kma ¼ 8), violent (kma ¼ 10), IPV (kma ¼ 2), and sexual specific counterpart (Sewall & Olver, 2019), while other lines
(kma ¼ 4) recidivism, institutional violence (kma ¼ 6), and of research have found that it did not increment the HCR-
institutional infractions in general (kma ¼ 5) (see Table 1). Of 20 (Douglas et al., 2005). Other lines of research (to be
note, given the debate on the predictive efficacy of the PCL-R discussed) further demonstrate that incremental predictive
for institutional violence (see DeMatteo et al., 2020; Hare et al., validity for sexual recidivism is weaker for the PCL-R rela-
2020; Olver et al., 2020), the criterion variable employed in tive to sexual offense risk measures (Sewall & Olver, 2019).
Table 1 is specifically acts of physical violence or aggression Finally, there have been other important indicators of cri-
where specified (i.e., in contrast to verbal abuse, other destruc- terion-related validity for PCL-R factors and facets. Factor 1,
tive behavior, or a general aggression variable that includes and notably the affective facet, have been identified as
both verbal and physical aggression). In contrast to Olver et al. responsivity issues as they correspond to treatment interfer-
(2020), we excluded meta-analyses that focused specifically on ing behaviors to be managed by service providers. These
Table 1. Summary of meta-analyses of the Hare PCL-R in the prediction of recidivism and institutional behavioral criteria.
Meta-analysis PCL-R variant Measure k n Metric ES d
General recidivism
Gendreau et al. (1996) PCL/R Total 9 1,040 r .29 .61
Salekin et al. (1996) PCL/R/SV Total 10 4,620 d .55 .55
Hemphill et al. (1998) PCL/R Total 7 1,275 r .27 .57
Walters (2003a) PCL/R/SV/YV Total 33 NR r .25 .52
Walters (2003b) PCL/R/SV/YV Factor 1 26 NR r .15 .30
Factor 2 26 NR r .32 .68
Hanson and Morton-Bourgon (2004) PCL-R/YV Total 9 1,966 d .67 .67
Leistico et al. (2008) PCL/R/SV/YV Total 62 11,140 d .50 .50
Factor 1 29 5,439 d .37 .37
Factor 2 29 5,439 d .64 .64
Meta-analytic kma Median d Muw d
aggregate Total 7 .55 .56
Factor 1 2 .34 .34
Factor 2 2 .66 .66
Violent recidivism
Salekin et al. (1996) PCL/R/SV Total 10 1,323 d .67 .67
Hemphill et al. (1998) PCL/R/SV Total 6 1,374 r .27 .57
Factor 2 27 NR r .26 .54
Hanson and Morton-Bourgon (2004) PCL-R/YV Total 9 2,446 d .58 .58
Campbell et al. (2009) PCL-R Total 24 4,757 r .27 .56
Yang et al. (2010) PCL-R Total 16 3,854 d .55 .55
PCL-R/SV Factor 1 13 3,895 d .22 .22
PCL-R/SV Factor 2 13 3,995 d .61 .61
PCL: SV Total 8 2,506 d .65 .65
Singh et al. (2011) PCL-R Total 20 3,854 DOR 2.08 .44
Hawes et al. (2013) PCL-R Total 13 3,467 d .55 .55
Factor 1 8 1,538 d .08 .08
Factor 2 8 1,454 d .63 .63
Mokros et al. (2014) PCL-R Total 6 1,562 AUC .708 .77
PCL: SV Total 2 340 AUC .633 .48
PCL-R/SV Total 7 1,806 AUC .698 .74
Factor 1 3 .22 .22
Factor 2 3 .63 .60
Intimate partner violence recidivism
Hanson et al. (2007) PCL-R Total 2 736 d .60 .60
van der Put et al. (2019) PCL-R/SV Total 4 908 AUC .68 .68
Sexual recidivism
Walters (2003b) PCL-R Factor 1 5 NR r .05 .10
Factor 2 5 NR r .08 .16
Hanson and Morton-Bourgon (2005) PCL-R/SV/YV Total 13 2,783 d .29 .29
Hawes et al. (2013) PCL-R Total 20 5,239 d .40 .40
Factor 1 13 2,838 d .17 .17
Factor 2 13 2,838 d .44 .44
Mokros et al. (2014) PCL-R Total 3 755 AUC .629 .47
PCL: SV Total 2 232 AUC .577 .28
PCL-R/SV Total 4 891 AUC .610 .40
Factor 1 2 .14 .14
Factor 2 2 .30 .30
Institutional misconduct
Walters (2003a) PCL/R/SV Total 15 NR r .27 .54
Factor 2 16 NR r .27 .54
Guy et al. (2005) PCL-R/SV Total 38 5,381 r .29 .61
Factor 1 25 3,219 r .21 .42
Factor 2 25 3,219 r .27 .54
Leistico et al. (2008) PCL-R/SV Total 45 6,137 d .53 .53
Factor 1 30 3,898 d .41 .41
Factor 2 29 3,848 d .51 .51
Olver et al. (2020) PCL-R/SV Total 5 320 d .35 .35
Factor 1 3 .41 .40
Factor 2 3 .54 .53
(continued)
Table 1. Continued.
Meta-analysis PCL-R variant Measure k n Metric ES d
Institutional violence
Salekin et al. (1996) PCL/SV Total 3 362 d .92 .92
Walters (2003b) PCL-R/SV/YV Factor 1 14 NR r .12 .24
Factor 2 14 NR r .22 .44
Guy et al. (2005) PCL-R/SV Total 22 3,502 r .17 .35
Factor 1 16 2,129 r .14 .28
Factor 2 16 2,129 r .15 .30
Campbell et al. (2009) PCL-R Total 5 626 r .14 .28
PCL: SV Total 7 504 r .22 .44
Hogan and Ennis (2010) PCL-R Total 3 254 r .21 .42
PCL: SV Total 8 827 r .26 .52
PCL-R/SV Total 12 1,313 r .26 .52
Olver et al. (2020) PCL-R/SV Total 9 1,350 d .39 .39
PCL-R/SV Factor 1 7 813 d .20 .20
PCL-R/SV Factor 2 7 813 d .52 .52
Factor 1 3 .24 .24
Factor 2 3 .44 .42
Note: PCL ¼ Psychopathy Checklist, PCL-R ¼ Psychopathy Checklist-Revised, PCL: SV ¼ Psychopathy Checklist: Screening Version; PCL: YV ¼ Psychopathy Checklist:
Youth Version. Most meta-analyses analyzed multiple variants of the PCL-R scales in aggregate and reported a single effect size. k ¼ number of studies, n ¼
participants, Metric ¼ type of metric employed in the original meta-analysis for reporting effect sizes; ES ¼ effect size originally reported; d ¼ standardized
mean difference (Cohen’s d) computed or directly reported from ES. When r (point biserial correlation) or area under the curve (AUC) was the original unit of
aggregation it was converted to d using formulae/tables from Rice and Harris (2005). DOR ¼ diagnostic odds ratio converted to d per Sanchez-Meca et al.
(2003). For Mokros et al. (2014), the AUC and SE from reported studies were used to compute AUCw and to convert to d via Rice and Harris (2005). NR ¼ not
reported. kma ¼ number of meta-analyses included in the meta-analytic aggregate; Median d ¼ median d value; Muw d ¼ mean unweighted d.
features are associated with weaker working alliances, par- obtained by independent raters is impacted by the context
ticularly the client-therapist bond (DeSorcy et al., 2020), (i.e., research vs. field context) of the ratings made.
decreased treatment progress (Olver et al., 2013; Sewall & The intraclass correlation coefficient (ICC) is used to
Olver, 2019), and increased program attrition (Olver & evaluate rater agreement. Consistency agreement (ICCC)
Wong, 2011). Blais et al. (2014) meta-analysis also found represents the rank ordering or linear association of two sets
Factor 2 to have stronger links to reactive (heated, impul- of ratings across the totality of cases (equivalent to a
sive) violence, and for Factor 1 to have prominent associa- Pearson r), while absolute agreement (ICCA) also takes into
tions with instrumental (goal driven, planned) violence. consideration the magnitude of disparity in scores between
raters on each case. Further, single measure ICCs are com-
puted based on independent ratings (ICCA1 or ICCC1), while
Interrater reliability
average measures ICC (ICCAk or ICCCk) represents the level
Evidence for the interrater reliability of PCL-R scores has of agreement after averaging scores between raters (McGraw
broadly come from two types of investigations. The first is & Wong, 1996). ICCA1 values are the most stringent metric
controlled studies featuring the interrater reliability of inde- for determining interrater reliability since they represent the
pendent ratings of the PCL-R in a research context by two level of agreement between independent sets of ratings
or more trained raters using comprehensive information across cases, the rank ordering of test scores, and the size of
sources such as detailed institutional files with or without a disparity in ratings. Several classification systems for inter-
clinical interview. If an interview is used, the participants preting ICC magnitudes have been proposed: Landis and
evaluated are aware that ratings are done for research pur- Koch (1977; < 0 Poor, 0-.20 Slight, .21-.40 Fair, .41-.60
poses only, information quality is generally high, and raters Moderate, .61-.80 Substantial, .81-1.00 Almost Perfect),
are well trained and have no obvious ax to grind. The Cicchetti and Sparrow (1981; < .40 Poor, .40-.59 Fair, .60-
second context features the interrater reliability of PCL-R .74 Good, .75-1.00 Excellent), Fleiss et al. (2003; < .40 Poor,
ratings obtained on the same client by independent evalua- .40-.74 Fair to Good, .75 to 1.00 Excellent), and Koo and Li
tors in real world field settings—known as field reliability— (2016; < .50 Poor, .50-.75 Moderate, .75-.90 Good, .90-
such as pretrial or presentence assessments, evaluation for 1.00 Excellent).
conditional release or correctional intake, or high stakes ICC values can be interpreted as representing the propor-
evaluations such as possible civil commitment pursuant to tion of variance in rater scores (e.g., PCL-R ratings) that
an SVP application. In such instances, information quality is reflects variation in levels of the trait measured (i.e., individ-
variable, evaluator training is uncontrolled, participants are ual differences in levels of psychopathy across the different
acutely aware they are being evaluated for decision making cases rated), while 1.00 – ICC represents the proportion of
purposes (which can influence their candor, information variability due to variability attributable to factors other
quality provided, or the decision to provide an interview), than levels of the trait being measured (e.g., rater error, bias,
and evaluators may have a “stake in game” depending on conscientiousness, information quality), collectively referred
whether they are completing an evaluation for the prosecu- to as error (Boccaccini et al., 2014). Invoking the most strin-
tion or defense. How “reliable” PCL-R scores are when gent system, such as Koo and Li (2016), to achieve
“Excellent” interrater reliability, the ICC value would repre- Minnesota Sex Offender Screening Tool-Revised (MnSOST-
sent at a minimum 90% of the variability in test scores R), and the PCL-R to be significantly different (ranging
being attributable to individual variability in the trait meas- from small to large, ds ¼ .34, .85, and .78, respectively)
ured, while only 10% of the variability would be attributable when scored by opposing evaluators, with the evaluators for
to error. If there is common ground among reliability classi- the Defense scoring lower than those for the State (Murrie
fication systems, it would be: a) higher levels of reliability et al., 2009); field interrater reliabilities were correspondingly
are desirable, and b) there is a wide threshold for tolerable ICCA1 ¼ .62 (27 pairs), .44 (27 pairs), and .42 (35 pairs),
levels of interrater agreement and variability due to error for respectively. Such allegiance effects can have significant
test scores to still be useful. implications for field reliability.
Interrater reliability for trained PCL-R raters with quality Rufino et al. (2012) interrater reliability study was a
information from clinical file with or without the interview hybrid examination of PCL-R ratings obtained under con-
tends to be quite high; the technical manual (Hare, 2003) trolled research versus field contexts; one of the conditions
cites ICCs of .86 (PCL-R total score), .75 (Factor 1), and .85 was an allegiance condition in which 44 Texas SVP evaluees
(Factor 2). Most studies regardless of context find interrater were rated independently on the PCL-R by evaluators
reliability to be somewhat lower for Factor 1 than Factor 2, retained by the State vs. the Defense (this was an expansion
presumably given the greater level of judgment involved in of the Murrie et al., 2009, sample). Under controlled
assessing, interpreting, and rating the interpersonal and research conditions, two trained graduate students inde-
affective features. Dåderman and Hellstr€ om (2018), in their pendently rating the 44 cases from the same file information
Swedish field simulation interrater reliability study, found available to the evaluators achieved excellent interrater reli-
strong interrater reliability by trained mental health profes- ability (PCL-R total ICCA1 ¼ .95, Factors 1 and 2 ICCA1 ¼
sionals across 43 pairs of PCL-R ratings of forensic psychi- .90), but the field reliability for PCL-R total scores between
atric patients (ICCA1 values PCL-R total ¼ .89, Factor 1 ¼ the State and Defense on the same 44 cases was poor
.82, Factor 2 ¼ .88). Further, in a large sample of PCL-R (ICCA1 ¼ .33). The interrater reliability of the graduate
trained raters (N ¼ 280) who had each rated a collection of research assistants research-based PCL-R ratings with the
six videotaped practice cases, Blais et al. (2017) found good State and Defense’s field ratings was even worse (ICCA1 ¼
reliability for PCL-R total (ICCA1 ¼ .75) and Factor 2 .29 and .14, respectively). Absent partisan bias, importantly,
(ICCA1 ¼ .78) scores, and moderate reliability for Factor 1 reliability can improve. An analogous Canadian study by
(ICCA1 ¼ .65). Harris et al. (2013), in a sample of court-remanded forensic
Several studies have reported high interrater reliability for inpatients, found that PCL-Rs rated from file by trained and
research based PCL-R ratings, for instance, with ICCA1 val- experienced research assistants had good interrater agree-
ues (PCL-R total score) of .81 (41 pairs, Dutch forensic ment with clinician field ratings of the PCL-R, rated inde-
inpatients, Zwets et al., 2015), .84 (32 pairs, Swedish prison pendently based file and interview (58 pairs, ICCA1 ¼ .79)
inmates; Skeem et al., 2007), .88 (51 pairs, court ordered as part of an assessment for adjudicative competence or
drug treatment attendees; Poythress et al., 2010), and .88 (45 criminal responsibility.
pairs, intellectually disabled UK forensic inpatients; With these considerations taken together, what does the
Morrissey et al., 2005). Moreover, good interrater reliability aggregate literature say about the state of field reliability
was found in a Canadian multisite sample of Indigenous with the PCL-R? In a PCL-R field reliability meta-analysis of
(117 pairs, ICCA1 PCL-R total ¼ .85, Factor 1 ¼ .74, Factor 13 studies Olver et al. (2020) found the overall ICC for
2 ¼ .82) and non-Indigenous (221 pairs, ICCA1 PCL-R total PCL-R total scores to be .68–.69 (k ¼ 10-13).2 As such,
¼ .87, Factor 1 ¼ .78, Factor 2 ¼ .84) federally sentenced across field settings, a little more than two-thirds of the vari-
men (Olver et al., 2018), as well as Latino (ICCA1 ¼ .89) ability in PCL-R scores could be attributed to trait variability
and African American (ICCA1 ¼ .83) U.S. male inmate sam- in psychopathy, and the remaining one-third or less would
ples (Sullivan et al., 2006)1. be due to non-trait related factors (e.g., evaluator allegiance,
When information quality is high, raters are trained and training, information quality). The ICC effect size, however,
well calibrated, and situational or contextual variables that was moderated by country. Across five Canadian studies,
which featured a combination of routine correctional set-
can contribute to error variance are absent, high interrater
tings and potential adversarial contexts (e.g., Dangerous
reliability can be achieved; when these conditions are absent,
Offender evaluations), the ICC effect sizes (k ¼ 3-4) ranged
not surprisingly, reliability plummets. The field setting can
from .78–.88 (the latter with outlier excluded), and on par
be an adversarial context, particularly with independent
with the ICC values reported in the PCL-R technical manual
evaluators retained by opposing counsel, which can also
(Hare, 2003). By contrast, six U.S. studies, all of which fea-
make field reliability a moving target. For instance, research
tured SVP evaluation (and hence, adversarial) contexts,
examining interrater agreement in the context of Texas SVP
interrater reliability effect sizes ranged from .56–.60 (k ¼ 5-
evaluations has demonstrated test scores on the Static-99,
6), representing “Fair to Good” or “Moderate” reliability
across classification systems.
1
As Sullivan et al. (2006) reported average measure ICC values (.94 and .91,
respectively) the Spearman Brown prophecy formula was used to convert
2
these to single measure estimates, as reported here. Olver et al. (2018) The k per effect varied depending on consideration for possible study overlap
reported ICCC1, but ICCA1 values are reported here (ICCC1 and ICCA1 values and the correlation metric used, 11 of which reported ICCA1, 1 reported the
were generally identical). equivalent of ICCC1, and 1 reported both.
Credibility in court: Putting psychometric findings predictor of sexual violence; it does not assess the dimension
into context of sexual deviance which is one of the central constructs in
sexual violence risk assessment, although the PCL-R cap-
What data are needed to support the credibility of the PCL- tures the general antisociality domain, another central con-
R in court? There are several implications from the psycho- struct. When high psychopathy and high sexual deviance are
metric and legal reviews covered thus far. First, there should present, the individual will likely be higher risk for future
be satisfactory evidence that the PCL-R measures a personal- sexual violence than when only one element is present, or
ity and psychopathology construct with criminal justice rele- neither are present.
vance that correlates with but can be distinguished from Third, the PCL-R should demonstrate evidence for satis-
parallel diagnoses. The previous review summarized evi- factory consistency in measurement when rated by inde-
dence for the factor structure of the PCL-R and more pendent evaluators. The previous review demonstrates that
broadly, the latent structure of psychopathic traits, and its PCL-R interrater reliability in controlled research contexts,
convergence with (and divergence from) clinically and the- field simulations, and routine field settings tends to be quite
oretically meaningful measures of symptomatology. strong; however, in adversarial contexts (e.g., SVP evalua-
Although debates may continue about whether antisociality tions), field reliability can be quite poor. So, how high is
is an inherent feature of the syndrome, the evidence is high enough? Our response is, the higher the better. For
robust supporting the construct validity of the PCL-R as a one, we know that adversarial allegiance can attenuate the
measure of psychopathy. interrater reliability of ratings on structured forensic meas-
Second, to be used in the context of risk assessment, the ures (even seemingly “objective” measures such as Static-99),
PCL-R would be expected to evince satisfactory predictive especially as seen in some high stakes legal proceedings.
validity for relevant antisocial behavioral criteria. None of Field reliability in routine correctional assessment contexts
the predictive validity ES magnitudes of the PCL-R for where adversarial allegiance may be less pronounced has
future antisocial behavior within the institution or the com- been higher for measures such as the Static-99 (Texas ICCA1
munity are large; the ES are medium for general, violent, ¼ .79, 600 pairs; New Jersey ICCA1 ¼ .88, 135 pairs,
and IPV recidivism and institutional behavior problems in Boccaccini, Murrie et al., 2012) as well as the PCL-R (e.g.,
general, the ES are small for sexual recidivism, and they are Ismail & Looman, 2018; Langton et al., 2006; Matsushima,
somewhere in the middle for institutional violence. This, 2016). Moreover, rater training has the potential to decrease
however, also characterizes the meta-analytic literature for measurement error and improve examiner agreement in rat-
purpose-built risk tools. For instance, formal meta-analytic ing the PCL-R (Boccaccini et al., 2014). Finally, PCL-R
comparisons demonstrate that PCL-R prediction magnitudes scores with modest absolute interrater agreement may still
for violent recidivism are not significantly different from be useful, although such a statement is not intended to jus-
those of the HCR-20, LSI measures, VRAG, or the VRS tify high levels of rater disagreement; higher reliability is
(Campbell et al., 2009; Gendreau et al., 2002; Yang et al., always preferred. Boccaccini et al. (2012) found Texas SVP
2010), nor were ES magnitudes different between the PCL-R PCL-R field reliability ratings across evaluations ranged
and these measures in the prediction of institutional vio- from ICCA1 .44 to .52 between State and Defense evaluators
lence (Campbell et al., 2009; Hogan & Ennis, 2010), or even across 38 pairs of cases; this would mean that 48% to 56%
when compared against IPV specific risk tools in the predic- of the variability in PCL-R ratings could be attributable to
tion of future IPV (Hanson et al., 2007; van der Put error or other factors unrelated to the trait measured
et al., 2019). (Boccaccini et al., 2014). However, two-thirds of the AUCs
The PCL-R, however, has significantly lower predictive were medium to large in magnitude in the prediction of
accuracy for sexual recidivism compared to sexual offense misconduct. Boccaccini et al. (2012) noted this to be because
specific tools demonstrated through meta-analysis. For both sets of evaluators ranked their evaluees similarly by
instance, the 95% CIs for PCL-R total score in the predic- PCL-R score, even though the defense tended to rate lower;
tion of sexual recidivism from Hawes et al. (2013) (d ¼ .40, as such, higher scores translated into an increased probabil-
.38, .53) do not overlap with the lower bounds of actuarial ity for misconduct on each side. This offers small comfort
(d ¼ .67, .63-.72) or mechanical (d ¼ .66, .58-.74) sexual for high stakes legal proceedings, however, and we concur
violence risk tools identified in the Hanson and Morton- with Boccaccini et al. (2012) that evaluators should prioritize
Bourgon (2009) meta-analysis, demonstrating that they improving their rating accuracy and taking steps to mitigate
come from different populations of ES. That said, the PCL- adversarial bias.
R does not fare any differently than general violence or gen-
eral risk-need tools in the prediction of sexual recidivism. A
corollary to this finding, however, is that the PCL-R has an Cross-examination on the PCL-R: Ten key questions
additive effect when combined with sexual deviance, either
Does the PCL-R speak to treatability?
assessed via a sexual violence risk tool, other structured
measure, or phallometrically. Hawes et al. (2013) found a No, research demonstrates that even high PCL-R scoring
combination of high PCL-R measured psychopathy com- individuals can make changes in treatment and do not inex-
bined with an indicator of sexual deviance equated to a 2.80 orably reoffend violently or otherwise (Olver et al., 2013;
to 3.21 increase in the odds of sexual recidivism. This likely Polaschek & Wong, 2020; Olver & Riemer, 2021; Sewall &
exists because the PCL-R on its own is not a strong Olver, 2019; Skeem et al., 2002; Wong et al., 2012). This
developing body of research (most of which features adult variables impinging on an individual that can affect the
male samples) shows that in RNR-based correctional pro- probability of a certain behavior at any given moment.
grams, high PCL-R scorers do not have a significantly differ-
ent amount of treatment change, posttreatment, from lower
scoring individuals. Per the RNR principles, high psychop- Is Factor 1 junk? Why do we need it? Are high total
athy persons stand to benefit most from high intensity, tar- scores with proportionately more Factor 1 items present
geted evidence-informed services and may be retained less predictive than those with proportionately more
in treatment. Factor 2 items present?
Factor 1 is not junk and it primarily represents a set of
responsivity issues (Wong, 2016), all of which can be man-
Does the PCL-R assess a static construct?
aged (Polaschek & Wong, 2020; Sewall & Olver, 2019;
No, although the PCL-R assesses a stable construct. There is Wong, 2016). Factor 1 is thus a signal for service providers
evidence from cross-sectional research for the stability of Factor and programs to take caution and to manage and contain
1 over age cohorts, while Factor 2 declines as evidenced by an potential treatment interfering behaviors in their clientele.
inverse association with age (Harpur & Hare, 1994; Olver & Meta-analyses, as reviewed previously, show that Factor 1
Wong, 2015). Moreover, most of the items from Factor 2, on its own predicts violence and other outcomes, particu-
including parasitic lifestyle, poor behavioral controls, lacks goals, larly the affective facet, but much of the association occurs
impulsivity, irresponsibility, and possibly, proneness to bore- by virtue of shared variance with Factor 2. There has been
dom, are routine targets for treatment in correctional programs little evidence to date to suggest, however, that PCL-R total
(e.g., cognitive problem-solving, anger management, vocational scores are more or less predictive depending on differential
retraining, educational upgrading, substance use treatment, weighting of Factor 1 and Factor 2 items; as an alternative,
alternatives to aggression). Further, some Factor 1 items are not we would recommend simply using a risk tool in combin-
immutable, such as does not accept responsibility, deceitfulness, ation with the PCL-R to establish risk formulations in add-
and manipulation, and are treated through programs for crim- ition to reporting PCL-R factor and facet scores. Beyond
inal attitudes/criminal thinking (e.g., increased perspective tak- any relevance to risk, the traits assessed by Factor 1 (e.g.,
ing, prosocial thinking). low empathy, remorselessness) are certainly relevant to clin-
ical and forensic case formulations.
Can the PCL-R assess change?
If Factor 2 is a better predictor of recidivism, especially
The PCL-R does not have a specific mechanism to assess
the antisocial facet, why do we need anything beyond
change; however, it should be periodically readministered
criminal history?
and rescored to evaluate possible changes that could occur
on any of the items or factor/facet dimensions over time or Factor 2 is not criminal history – it is a set of behavioral
in response to treatment or other change agents. indicators of psychopathy, some of which include a record
of antisocial behavior, which can increase for the worse, or
improve for the better with time and appropriate interven-
Does the PCL-R predict all forensic behavioral
tion. The antisocial facet, specifically, is a latent antisociality
criteria well?
construct that reflects a propensity for antisocial behavior,
The PCL-R has satisfactory predictive validity for future vio- regardless of whether the behavior is formally adjudicated
lence, domestic violence, breaches, general/nonviolent reof- and criminal in nature (Hare & Neumann, 2010). With that
fending, and institutional conduct problems; it is a weaker said, criminal history is critical to obtaining high scores on
predictor (small effect) of sexual recidivism. Satisfactory pre- several PCL-R items, including Items 18 (juvenile delin-
diction of highly specific and extremely low base rate criteria quency), 19 (revocation of conditional release), and 20
(e.g., homicide recidivism) has not been demonstrated for the (criminal versatility); these are static risk factors. Items 10
PCL-R or any measure. Not all high PCL-R scorers are high (poor behavioral controls) and 12 (early behavioral prob-
risk for sexual violence, especially absent any history of sexual lems) either reflect dynamic risk factors and/or are not
offending. Further, not all high PCL-R scorers are necessarily necessarily specific to criminal behavior.
high risk for general violence, intimate partner violence, or
institutional violence. The PCL-R needs to be paired with
another tool to make that kind of determination. How can you justify making categorical determinations
of a dimensional construct, such as the 30-point
cut score?
Can the PCL-R definitively predict if someone will
Given the dimensionality of the PCL-R and the psychopathy
reoffend violently or otherwise?
construct (e.g., Edens et al., 2006), we recommend also using
No measure can definitively predict whether an individual percentile ranks when reporting an individual’s level of
will reoffend in any capacity, given the complexity of human psychopathy as well as that of their factor/facet profiles to
behavior and the enormous volume of unknown situational, the appropriate norm group (e.g., correctional, psychiatric,
contextual, personal, interpersonal, and environmental male vs. female). The PCL-R manual (Hare, 2003) notes that
there is an SEM of around three or four points, which magnitudes of .85 or higher (e.g., Rufino et al., 2012; Sewall
makes using a definitive cut score problematic. For instance, & Olver, 2019), there is little reason for trained professionals
there is little difference between a score of 28 or 29 versus a not to be able to obtain this threshold.
score of 30. Using percentile ranks and standard deviation
units, however, is one means of addressing the problem. For
instance, it is telling if an individual is at the 90th percentile Admissibility of the PCL-R
and 1.5 SDs above the mean for a typical North American The data from clinician surveys and case law reviews suggest
male in a correctional setting; high scores can be a concern, the PCL-R is being used with increasing frequency in a var-
and the question reframed could be, how psychopathic is this iety of legal contexts. Despite this increased use, data suggest
individual? as opposed to, does this individual have psychop- the introduction of the PCL-R is rarely challenged by oppos-
athy? That said, we recommend this not be taken in a vac- ing parties. It is important, however, not to equate the low
uum. Consider the scenario in which an individual with a base rate of PCL-R admissibility challenges with the idea
PCL-R total score of 32 is engaging well in a violence reduc- that the PCL-R is always being introduced in legal contexts
tion program, retaining a job in the greenhouse, has not had for which there is empirical support for its use. Case law
an institutional infraction in 3 months, and is setting a posi- reviews found that some mental health professionals were
tive example on the unit and in group. using the PCL-R in contexts for which there is little empir-
ical support (see DeMatteo et al., 2014b; DeMatteo & Edens,
How high does predictive validity need to be? 2006), and it is likely that some attorneys do not have suffi-
cient knowledge about appropriate uses of the PCL-R to
Predictive validity needs to be high enough to be useful. We permit them to raise a cogent challenge to its admissibility
would suggest at least moderate accuracy (AUCs mid .60 s and in certain legal contexts.
up) is desirable, as this enables a decision maker to know that As previously discussed, PCL-R scores are a valid pre-
people with high scores on the tool—taken together with other dictor of several outcomes that are interest to the criminal
assessment and collateral information—will require special justice system. However, validity is not a static property of a
management considerations. It is also important to bear in psychological test (Messick, 1995), so it would be incorrect
mind that even small effects can be useful depending on the to assert that the PCL-R is “valid” without appropriately
purpose (re: Rosenthal’s 1990 aspirin effect), but it speaks to qualifying and contextualizing that assertion. A test is valid
the need for an evaluation to be buttressed with even more for particular purposes with particular populations.
information (e.g., sexual violence risk assessment). Ultimately, Therefore, in legal contexts, the PCL-R is valid only in par-
risk assessment is about management. Let us expand this ques- ticular court proceedings, for particular legally relevant ques-
tion to ask, what if, hypothetically, the PCL-R was 100% pre- tions, and when the PCL-R results are communicated in
dictive? (e.g., cut score of say, 33 and AUC ¼ 1.0)—that this particular ways. Although test validity is not the sole criter-
person will be violent in the general population toward staff or ion used by courts when they are asked to determine the
other inmates without action taken to manage risk. What admissibility of a psychological test, it is a core component
would we do? In such instances, we would manage the risk. of the trial judge’s gatekeeping analysis of admissibility.
And to manage risk means that risk is not static, PCL-R score In United States federal courts, evidentiary admissibility
notwithstanding. Being managed means that risk is being is governed by the Federal Rules of Evidence (FRE; 1975).
handled in a way to promote safety of staff, inmates/patients, State courts can use their own rules of evidence, but nearly
and the clients themselves. Per the RNR principles this could every state has adopted the FRE in whole, in part, or with
include treatment programming, higher levels of security, staff minor modifications (Weinstein & Berger, 2015). The FRE
surveillance, core correctional engagement practices, and super- apply to all federal civil cases and proceedings, criminal
vised contact with staff and inmates/patients. In turn, the level cases and proceedings, and contempt proceedings (unless
of security can be titrated and restrictions lifted via attention the court is acting summarily), but they do not apply to a
to dynamic security, monitoring offense linked proxy behav- court’s determination of a preliminary question of fact gov-
iors, and increasing new prosocial behaviors. erning admissibility, grand jury proceedings, and several
other proceedings (e.g., extradition, sentencing, granting/
revoking probation or supervised release, whether to release
How high does interrater reliability need to be?
an arrestee on bail) (see Federal Rule of Evidence 1101).
At a minimum, interrater reliability needs to be high enough The admissibility of all evidence, including lay/fact evidence
for test scores to be valid. Although scores based on samples and expert evidence, is initially governed by FRE 401 (Test for
and contexts that generate modest reliabilities have gener- Relevant Evidence) and FRE 403 (Excluding Relevant Evidence
ated acceptable prediction magnitudes for forensic behav- for Prejudice, Confusion, Waste of Time, or Other Reasons).
ioral criteria (Boccaccini et al., 2012), higher reliability Per FRE 401, for evidence to be admissible, it must be relevant,
amounts to more useful and trustworthy information. We which means it has probative value in the matter before the
need to do better. Trained raters with sufficient information court. But proffered evidence that is relevant can nevertheless
and with preconceived biases in check should be able to get be excluded under FRE 403 if, for example, the evidence
within 1.0 SEM (e.g., four or so points) on the PCL-R total would be unfairly prejudicial or misleading to the jury.
score. Further, if student raters routinely achieve ICCA1 Expert evidence, which would be the type of evidence in
which the PCL-R would be discussed, is further governed by in the community is an important consideration. Such con-
FRE 702 (Testimony by Expert Witnesses). texts could include determining whether an offender should
In addition to court rules, there are several court cases be released on parole or whether an offender should receive
that govern the admissibility of expert evidence: Frye v. community supervision (e.g., probation) or a custodial sen-
United States (1923) (establishing “general acceptance” test), tence. PCL-R evidence could also be probative in determin-
Daubert v. Merrell Dow Pharmaceuticals (1993) (holding ing whether a convicted offender should receive a shorter or
Frye was superseded by the FRE for determining the admis- longer sentence of incarceration, given that a shorter sen-
sibility of expert evidence in federal courts),3 and Kumho v. tence means the offender would be in the commu-
Carmichael Tire Co. (1999) (holding Daubert applies to all nity sooner.
expert evidence). To be clear, although there is some empirical support for
The 401/403 admissibility analysis as applied to PCL-R the association of PCL-R scores with future community vio-
evidence will vary depending on the legal context because lence, the available research suggests the relationship is
the outcome of interest and, importantly, the PCL-R’s ability moderate on average (e.g., Gendreau et al., 2002; Leistico
to offer probative value vis-a-vis the outcome varies by con- et al., 2008; Olver & Wong, 2015). Further, the PCL meas-
text. For example, the PCL-R will have more probative value ures in isolation are comparable but not better predictors of
in legal contexts in which it is being used in a manner that reoffending relative to risk assessment measures (Campbell
is empirically supported, and it will have correspondingly et al., 2009; Gendreau et al., 2002; Singh et al., 2011; Yang
less probative value if it is being used to assist courts in et al., 2010). Nevertheless, the bar for concluding that evi-
addressing novel legal issues for which the PCL-R has not dence is relevant/probative (FRE 401) is relatively low, while
been evaluated empirically. As such, it follows that it is pos- the bar for concluding that relevant evidence is prejudicial
sible for PCL-R evidence to be probative in some types of or misleading (FRE 403) is fairly high. Given moderate rela-
legal proceedings but not in others, and it also possible for tionship of the PCL-R scores with community recidivism,
PCL-R evidence to be overly prejudicial or misleading in and the mixed evidence for a prejudicial labeling effect
some contexts but not in others. when offenders are referred to as “psychopaths,” it is likely
Although a full discussion of the admissibility of the that the PCL-R would be admissible in contexts in which
PCL-R across multiple legal contexts is beyond the scope of community recidivism is an important consideration.
this paper, below we offer some thoughts on the PCL-R’s
admissibility in several legal contexts in which it is com-
monly used.4 (For a discussion of whether PCL-R evidence Sexually violent predator (SVP) commitment
satisfies the admissibility standards in other legal contexts, Case law reviews and clinician surveys suggest that the PCL-
readers are referred to DeMatteo et al., 2016, 2019). Using R is used more frequently in SVP hearings than in any other
the 401/403 admissibility framework, the key question is type of legal proceeding (DeMatteo et al., 2014a, DeMatteo
whether PCL-R evidence is both probative and not unduly et al., 2014b; Jackson & Hess, 2007; Walsh & Walsh, 2006).
prejudicial or misleading. We recommend that readers con- The key legal question in SVP contexts is whether the
sult Table 1 while reading the next three sections. offender poses a significant risk of engaging in future sexu-
ally violent behavior. There is mixed evidence regarding the
Recidivism in the community PCL-R’s ability to predict sexual violence. Some studies have
found that PCL-R scores are poor predictors of sexual reof-
Several meta-analyses have examined the predictive accuracy fending (e.g., Eher et al., 2015; Murrie et al., 2012, whereas
of PCL-R scores for criminal recidivism in the community others have concluded that PCL scores are relevant to recid-
(see DeMatteo et al., 2021; Douglas et al., 2018, for reviews). ivism risk (e.g., Hawes et al., 2013). There is also some
Given the PCL-R’s ability to predict community recidivism, meta-analytic evidence that the PCL-R’s predictive effects
PCL-R evidence would likely be viewed as probative if it is were stronger for scores calculated for research rather than
introduced in legal contexts in which an offender’s behavior clinical use (Hawes et al., 2013). The ability of the PCL-R to
predict both short-term and longer-term sexual violence
3
Daubert is the admissibility standard in all federal courts, and most states recidivism is also mixed, although there is more evidence
have adopted Daubert or a close derivative (see Slobogin et al., 2020). supporting the PCL-R’s predictive validity over longer peri-
4
Because a 702 analysis is dependent on the specific expert and the specific
nature of the proffered expert evidence in the context of a specific case, we
ods of time. The weight of the evidence suggests PCL-R
will limit our discussion to whether the PCL-R satisfies the 401/403 scores are meaningfully related to long-term sexual recidiv-
admissibility standard in several legal contexts. With that said, we offer the ism (e.g., Hildebrand et al., 2004; Parent et al., 2011; Rice &
following brief comment regarding the PCL-R and Daubert. If a court
determines that PCL-R evidence is relevant (FRE 401) and not unduly Harris, 1997), particularly if they are considered in combin-
prejudicial (FRE 403) in a specific case, we believe the PCL-R could satisfy the ation with deviant sexual arousal (Hawes et al., 2013;
four Daubert criteria: testing (PCL-R has been subjected to rigorous empirical Hildebrand et al., 2004; Olver & Wong, 2006).
analysis in several legal contexts), peer review and publication (well-developed
empirical literature regarding the use of the PCL-R in multiple contexts), error There is also mixed evidence regarding the potential
rate (discussion of SEM established via meta-analyses and meta-meta- prejudicial impact of psychopathy evidence in SVP contexts.
analyses), and general acceptance (use of the PCL-R is generally accepted In early research, Guy and Edens (2003) found that SVP
among practitioners in several contexts). Again, though, we emphasize that
the admissibility of expert evidence cannot be discussed in a vacuum because commitment decisions (examined among undergraduate
it is dependent on the specifics of each case. research participants) were not influenced by the
psychopathy label among male participants, but that female articles recently published by two groups of authors who
participants were more likely to commit those described by have taken opposing views on this issue (see DeMatteo
the prosecution as psychopathic. In another study, et al., 2020; Olver et al., 2020). Based on their view of the
Lieberman and Krauss (2009) compared biases of the psych- empirical literature, DeMatteo et al. (2020) concluded that
opathy label to another diagnosis in SVP hearings. In that “the overall association between PCL-R scores and violence
study, undergraduates were asked to make a commitment at the group level is only moderate in terms of effect size,
decision for a male offender diagnosed as a “psychopath” or both in absolute terms and relative to the effect size of other
with a paraphilia who had an extensive sexual offense his- established risk factors for violence; the association between
tory and victimized children of both sexes. They found that PCL-R scores and violence in institutional settings is small
the psychopathy diagnosis created bias and that all partici- in terms of effect size; and the association between PCL-R
pants were more likely to support SVP commitment. scores and serious institutional violence is negligible” (p.
Taken together, there is ambiguous evidence regarding 137). DeMatteo et al. (2020) also concluded that “one can-
the PCL-R’s ability to predict shorter-term sexual recidivism, not use the PCL-R in the context of capital sentencing eval-
which raises concerns about its relevance and probative uations to make predictions that an individual will engage
value (FRE 401) in those contexts. However, there is some in serious violence in high-security institutional settings
supporting evidence for using the PCL-R to predict longer- with adequate precision or accuracy to justify reliance on
term sexual recidivism, although only a few studies have the PCL-R scores” (p. 137). By contrast, Olver et al. (2020)
examined follow-up periods in excess of 10 years. Evidence concluded that “the PCL scales demonstrate predictive valid-
for prejudicial labeling effects (FRE 403) is mixed and likely ity for institutional violence, including ‘serious’ violence,
would not rise to the level of calling for exclusion of the and do so with robustness (i.e., medium in ES magnitude),
PCL-R in SVP contexts. Given evidence of some probative comparable with other tools, including those designed to
value and inconclusive evidence of prejudicial impact, the assess risk for violence or different outcomes” (p. 498).
PCL-R is likely to be admitted for purposes of long-term Based on their review of the literature, Olver et al. (2020)
SVP commitment considerations. recommended that “the PCL-R (and its derivatives) can and
should be part of a comprehensive violence risk assessment”
(p. 506).
Capital sentencing This debate about the PCL-R’s predictive validity in insti-
In its reinstatement of the death penalty in the United States tutional settings is directly relevant to the FRE 401/403 ana-
in 1976, the Supreme Court of the United States held that lysis regarding the admissibility of the PCL-R in capital
sentencing juries in capital cases must be given guidance in contexts. If one concludes that the PCL-R’s ability to predict
deciding whether death is an appropriate punishment (e.g., serious institutional violence – the type of violence that
Gregg v. Georgia, 1976). When deciding on an appropriate would likely be considered by a capital sentencing jury
punishment after an offender has been convicted of a capital when focusing on “future dangerousness” – is relatively
offense, sentencing juries consider aggravating factors weak, then introduction of the PCL-R would not satisfy FRE
(which support the imposition of the death penalty) and 401 because it would not be probative to the legal issue. If,
mitigating factors (which support the imposition of life in however, one concludes that the PCL-R can predict serious
prison). One aggravating factor outlined by some states that institutional violence with sufficient accuracy, then the PCL-
has relevance to the PCL-R is a capital defendant’s risk of R would be viewed as relevant/probative and therefore
future danger (Fairfax-Columbo & DeMatteo, 2017), which admissible in capital sentencing contexts.
is the probability that an individual, absent imposition of
the death penalty, will engage in future violent behavior.
Final conclusions
Currently, most death penalty states require or permit con-
sideration of future dangerousness by capital sentencing Use of the PCL-R in legal contexts has been increasing over
juries (see Cunningham & Goldstein, 2003; DeMatteo the past 15þ years, and there is no reason to think that its
et al., 2020). use will decline. As such, it is important that forensic mental
In capital sentencing contexts, the issue of future danger- health professionals, legal practitioners, courts, and correc-
ousness should be interpreted in an appropriate context – tional personnel have a solid understanding of the PCL-R’s
i.e., whether the offender will be violent while incarcerated strengths and limitations, along with a clear understanding
in a high-security correctional facility (e.g., Cunningham, of appropriate uses, inappropriate uses, and best practices.
2006, 2008; Edens et al., 2005); in most cases, release to the As with any assessment tool, the PCL-R provides useful and
community is not an option. To be clear, although many defensible information in some contexts, but less so in other
courts have recognized that violence risk assessments in cap- contexts; the utility of the PCL-R is context-dependent, and
ital cases are specific to the correctional context (e.g., United knowing when and how to rely on the PCL-R is an essential
States v. Sablan, 2006), some courts have taken a broader component of responsible practice.
approach and considered a capital defendant’s “danger to In this article, we reviewed accepted uses of the PCL-R;
society” (e.g., Coble v. Texas, 2010). frequency of use of the PCL-R in court; the current state of
There is robust disagreement regarding the PCL-R’s abil- scientific knowledge regarding the PCL-R’s construct valid-
ity to predict institutional violence, as reflected in a series of ity, predictive validity, and interrater reliability; and the
credibility and admissibility of the PCL-R in various legal Cooke, D. J., & Michie, C. (2001). Refining the construct of psychop-
proceedings. We hope this comprehensive review will pro- athy: Towards a hierarchical model. Psychological Assessment, 13(2),
171–188. https://doi.org/10.1037//1040-3590.13.2.171
vide those who research, use, and rely on the PCL-R with a Cunningham, M. D. (2006). Dangerousness and death: A nexus in
useful summary and framework for using the PCL-R in legal search of science and reason. The American Psychologist, 61(8),
proceedings. As with all science, the take-home points and 828–839. https://doi.org/10.1037/0003-066X.61.8.828
conclusions regarding the PCL-R are subject to change as Cunningham, M. D. (2008). Forensic psychology evaluations at capital
new data accumulate, so we hope this article will also stimu- sentencing. In R. Jackson (Ed.), Learning forensic assessment (pp.
211–238). Routledge/Taylor & Francis Group.
late continued research on the PCL-R’s relevance and utility Cunningham, M. D., & Goldstein, A. M. (2003). Sentencing determina-
in legal contexts, particularly in areas and with populations tions in death penalty cases. In A. M. Goldstein & I. B. Weiner
for which we are lacking sufficient data. (Eds.), Handbook of psychology: Volume 11 – Forensic psychology
(pp. 407–436). John Wiley & Sons.
Dåderman, A. M., & Hellstr€om, A. (2018). Interrater reliability of
ORCID Psychopathy Checklist- Revised: Results on multiple analysis levels
for a sample of patients undergoing forensic psychiatric evaluation.
David DeMatteo http://orcid.org/0000-0002-0266-5749 Criminal Justice and Behavior, 45(2), 234–263. https://doi.org/10.
Mark E. Olver http://orcid.org/0000-0002-2265-0034 1177/0093854817747647
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993).
de Vogel, V., de Ruiter, C., Bouman, Y. H. A., & de Vries Robbe, M.
References (2009). SAPROF: Guidelines for the assessment of protective factors
for violence risk. In Forum Educatief.
Andrews, D. A., & Bonta, J. (1994–2010). The psychology of criminal
DeMatteo, D., & Edens, J. F. (2006). The role and relevance of the
conduct (1st to 5th ed.). Anderson.
Psychopathy Checklist-Revised in court: A case law survey of U.S.
Andrews, D. A., & Bonta, J. (1995). LSI-R: The level of service inven-
courts (1991-2004). Psychology, Public Policy, and Law, 12(2),
tory-revised. Multi-Health Systems. 214–241. https://doi.org/10.1037/1076-8971.12.2.214
Andrews, D. A., Bonta, J., & Wormith, J. S. (2004). Level of Service/ DeMatteo, D., Edens, J. F., Galloway, M., Cox, J., Smith, S. T., &
Case Management Inventory (LS/CMI): An offender assessment sys- Formon, D. (2014a). The role and reliability of the Psychopathy
tem. User’s guide. Multi-Health Systems. Checklist-Revised in U.S. sexually violent predator evaluations: A
Babchishin, K. M., & Helmus, L. M. (2016). The influence of base rates case law survey. Law and Human Behavior, 38(3), 248–255. https://
on correlations: An evaluation of proposed alternative effect sizes doi.org/10.1037/lhb0000059
with real-world data. Behavior Research Methods, 48(3), 1021–1031. DeMatteo, D., Edens, J. F., Galloway, M., Cox, J., Smith, S. T., Koller,
https://doi.org/10.3758/s13428-015-0627-7 J. P., & Bersoff, B. (2014b). Investigating the role of the Psychopathy
Blais, J., Forth, A. E., & Hare, R. D. (2017). Examining the interrater Checklist-Revised in United States case law. Psychology, Public
reliability of the Hare Psychopathy Checklist-Revised across a large Policy, and Law, 20(1), 96–107. https://doi.org/10.1037/a0035452
sample of trained raters. Psychological Assessment, 29(6), 762–775. DeMatteo, D., Haney-Caron, E., Mowle, E., & Edens, J. F. (2021). The
https://doi.org/10.1037/pas0000455 use of measures of psychopathy in violence risk assessment. In K. S.
Blais, J., Solodukhin, E., & Forth, A. E. (2014). A meta-analysis explor- Douglas & R. K. Otto (Eds.), Handbook of violence risk assessment
ing the relationship between psychopathy and instrumental versus (2nd ed., pp. 28–51). Routledge/Taylor & Francis Group.
reactive violence. Criminal Justice and Behavior, 41(7), 797–821. DeMatteo, D., Hart, S. D., Heilbrun, K., Boccaccini, M. T.,
https://doi.org/10.1177/0093854813519629 Cunningham, M. D., Douglas, K. S., Dvoskin, J. A., Edens, J. F.,
Boccaccini, M. T., Murrie, D. C., Rufino, K. A., & Gardner, B. O. Guy, L. S., Murrie, D. C., Otto, R. K., Packer, I. K., & Reidy, T. J.
(2014). Evaluator differences in Psychopathy Checklist-Revised fac- (2020). Statement of concerned experts on the use of the Hare
tor and facet scores. Law and Human Behavior, 38(4), 337–345. Psychopathy Checklist-Revised in capital sentencing to assess risk
https://doi.org/10.1037/lhb0000069 for institutional violence. Psychology, Public Policy, and Law, 26(2),
Boccaccini, M. T., Turner, D. B., & Murrie, D. C. (2008). Do some 133–144. https://doi.org/10.1037/law0000223
evaluators report consistently higher or lower PCL-R scores than DeMatteo, D., Hodges, H., & Fairfax-Columbo, J. (2016). An examin-
others? Findings from a statewide sample of sexually violent preda- ation of whether Psychopathy Checklist-Revised (PCL-R) evidence
tor evaluations. Psychology, Public Policy, and Law, 14(4), 262–283. satisfies the relevance/prejudice admissibility standard. In B. H.
Bornstein & M. K. Miller (Eds.), Advances in psychology and law
https://doi.org/10.1037/a0014523
(Vol. 2, pp. 205–239). Springer Publishing.
Boccaccini, M. T., Turner, D. B., Murrie, D. C., & Rufino, K. A.
DeMatteo, D., Murrie, D. C., Edens, J. F., & Lankford, C. (2019).
(2012). Do PCL-R scores from state or defense experts best predict
Psychopathy in the courts. In M. DeLisi (Ed.), Routledge inter-
future misconduct among civilly committed sex offenders? Law and
national handbook of psychopathy and crime (pp. 645–664).
Human Behavior, 36(3), 159–169. https://doi.org/10.1037/h0093949 Routledge/Taylor & Francis Group.
Boer, D. P., Hart, S. D., Kropp, P. R., & Webster, C. D. (1997). DeSorcy, D. R., Olver, M. E., & Wormith, J. S. (2020). Working alli-
Manual for the Sexual Violence Risk-20: Professional guidelines for ance and psychopathy: Linkages to treatment outcome in a sample
assessing risk of sexual violence. Institute Against Family Violence. of treated sexual offenders. Journal of Interpersonal Violence, 35(7-
Bonta, J., & Andrews, D. A. (2017). The psychology of criminal conduct 8), 1739–1760. https://doi.org/10.1177/0886260517698822
(6th ed.). Anderson. Douglas, K. S., Hart, S. D., Webster, C. D., & Belfrage, H. (2013).
Campbell, M. A., French, S., & Gendreau, P. (2009). The prediction of HCR-20V3: Assessing risk for violence–User guide. Mental Health,
violence in adult offenders: A meta-analytic comparison of instru- Law, and Policy Institute, Simon Fraser University.
ments and methods of assessment. Criminal Justice and Behavior, Douglas, K. S., Vincent, G. M., & Edens, J. F. (2018). Risk for criminal
36(6), 567–590. https://doi.org/10.1177/0093854809333610 recidivism: The role of psychopathy. In C. J. Patrick (Ed.),
Cicchetti, D. V., & Sparrow, S. A. (1981). Developing criteria for estab- Handbook of psychopathy (2nd ed., pp. 682–709). Guilford.
lishing interrater reliability of specific items: Applications to assess- Douglas, K. S., Yeomans, M., & Boer, D. P. (2005). Comparative valid-
ment of adaptive behavior. American Journal of Mental Deficiency, ity analysis of multiple measures of violence risk in a sample of
86(2), 127–137. criminal offenders. Criminal Justice and Behavior, 32(5), 479–510.
Coble v. Texas, 330 S.W.3d 253 (Tx. Ct. App. 2010). https://doi.org/10.1177/0093854805278411
Edens, J. F., & Cox, J. (2012). Examining the prevalence, role and impact (Corrections Research User Report No. 2007–07). Public Safety
of evidence regarding Antisocial Personality, sociopathy and psychop- Canada.
athy in capital cases: a survey of defense team members. Behavioral Hare, R. D. (1980). A research scale for the assessment of psychopathy
Sciences & the Law, 30(3), 239–255. https://doi.org/10.1002/bsl.2009 in criminal populations. Personality and Individual Differences, 1(2),
Edens, J. F., Buffington-Vollum, J. K., Keilen, A., Roskamp, P., & 111–119. https://doi.org/10.1016/0191-8869(80)90028-8
Anthony, C. (2005). Predictions of future dangerousness in capital Hare, R. D. (1991). Manual for the Revised Psychopathy Checklist.
murder trials: Is it time to “disinvent the wheel"? Law and Human Multi-Health Systems.
Behavior, 29(1), 55–86. https://doi.org/10.1007/s10979-005-1399-x Hare, R. D. (2003). The Hare Psychopathy Checklist-Revised (2nd ed.).
Edens, J. F., Hart, S. D., Johnson, D. W., Johnson, J. K., & Olver, M. E. Multi-Health Systems.
(2000). Use of the Personality Assessment Inventory to assess psych- Hare, R. D. (2016). Psychopathy, the PCL-R, and criminal Justice: Some
opathy in offender populations. Psychological Assessment, 12(2), new findings and current issues. Canadian Psychology/Psychologie
132–139. https://doi.org/10.1037//1040-3590.12.2.132 Canadienne, 57(1), 21–34. https://doi.org/10.1037/cap0000041
Edens, J. F., Marcus, D. K., Lilienfeld, S. O., & Poythress, N. G. (2006). Hare, R. D., & Neumann, C. S. (2010). The role of antisociality in the psych-
Psychopathic, not psychopath: Taxometric evidence for the dimen- opathy construct: Comment on Skeem and Cooke (2010). Psychological
sional structure of psychopathy. Journal of Abnormal Psychology, Assessment, 22(2), 446–454. https://doi.org/10.1037/a0013635
115(1), 131–144. https://doi.org/10.1037/0021-843X.115.1.131 Hare, R. D., Olver, M. E., Stockdale, K. C., Neumann, C. S., Mokros,
Eher, R., Olver, M. E., Heurix, I., Schilling, F., & Rettenberger, M. A., Baskin-Sommers, A., Brand, E., Folino, J., Gacono, C., Gray,
(2015). Predicting reoffense in pedophilic child molesters by clinical N. S., Kiehl, K., Knight, R., Leon-Mayer, E., Logan, M., Meloy, J. R.,
diagnoses and risk assessment. Law and Human Behavior, 39(6), Roy, S., Salekin, R. T., Snowden, R., Thomson, N., … Yoon, D.
571–580. https://doi.org/10.1037/lhb0000144 (2020). The PCL-R and capital sentencing: Reply to DeMatteo et al.
Fairfax-Columbo, J., & DeMatteo, D. (2017). Reducing the dangers of (2020b). Psychology, Public Policy, and Law, 26(4), 519–522. https://
future dangerousness testimony: Applying the Federal Rules of doi.org/10.1037/law0000290
Evidence to capital sentencing. William and Mary Bill of Rights Harpur, T. J., & Hare, R. D. (1994). Assessment of psychopathy as a
Journal, 25(3), 1047–1072. function of age. Journal of Abnormal Psychology, 103(4), 604–609.
Federal Rules of Evidence for United States Courts and Magistrates. https://doi.org/10.1037/0021-843X.103.4.604
(1975). West Publishing Company. Harpur, T., J., Hare, R. D., & Hakstian, A. R. (1989). Two-factor concep-
Fleiss, J. L., Levin, B., & Paik, M. C. (2003). Statistical methods for rates tualization of psychopathy: Construct validity and assessment implica-
tions. Psychological Assessment: A Journal of Consulting and Clinical
and proportions (3rd ed.). Wiley.
Psychology, 1(1), 6–17. https://doi.org/10.1037/1040-3590.1.1.6
Forth, A. E., Kosson, D. S., & Hare, R. D. (2003). Psychopathy checklist:
Harris, G. T., Rice, M. E., & Cormier, C. A. (2013). Research and clin-
Youth version (PCL: YV). Multi-Health Systems.
ical scoring of the Psychopathy Checklist can show good agreement.
Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).
Criminal Justice and Behavior, 40(11), 1349–1362. https://doi.org10.
Gagnon, N., Douglas, K., & DeMatteo, D. (2007, June). The introduc-
1177/0093854813492959 https://doi.org/10.1177/0093854813492959
tion of the Psychopathy Checklist-Revised in Canadian courts: Uses
Hart, S. D., & Hare, R. D. (1989). Discriminant validity of the
and misuses [Paper presentation]. Paper Presented at the 7th Annual
Psychopathy Checklist-Revised in a forensic psychiatric population.
Conference of the International Association of Forensic Mental
Psychological Assessment: A Journal of Consulting and Clinical
Health Services, Montreal, Quebec, Canada.
Psychology, 1(3), 211–218. https://doi.org/10.1037/1040-3590.1.3.211
Gardner, B. O., & Boccaccini, M. T. (2017). Does the convergent valid-
Hart, S. D., Cox, D. N., & Hare, R. D. (1995). The Hare Psychopathy
ity of the PAI Antisocial Features Scale depend on offender response
Checklist: Screening version (Pcl: Sv). Multi-Health Systems.
style? Journal of Personality Assessment, 99(5), 481–493. https://doi. Hart, S. D., Kropp, P. R., & Hare, R. D. (1988). Performance of psy-
org/10.1080/00223891.2017.1296846 chopaths following conditional release from prison. Journal of
Gendreau, P., Goggin, C., & Smith, P. (2002). Is the PCL-R really the Consulting and Clinical Psychology, 56(2), 227–232. https://doi.org/
“unparalleled” measure of offender risk? A lesson in knowledge 10.1037/0022-006X.56.2.227
cumulation. Criminal Justice and Behavior, 29(4), 397–426. https:// Hawes, S. W., Boccaccini, M. T., & Murrie, D. C. (2013). Psychopathy
doi.org/10.1177/0093854802029004004 and the combination of psychopathy and sexual deviance as predic-
Gendreau, P., Little, T., & Goggin, C. (1996). A meta-analysis of the tors of sexual recidivism: Meta-analytic findings using the
predictors of adult offender recidivism: What works! Criminology, Psychopathy Checklist-Revised. Psychological Assessment, 25(1),
34(4), 575–595. https://doi.org/10.1111/j.1745-9125.1996.tb01220.x 233–243. https://doi.org/10.1037/a0030391
Gregg v. Georgia, 428 U.S. 153 (1976). Helmus, L., Thornton, D., Hanson, R. K., & Babchishin, K. M. (2012).
Guy, L. S., & Edens, J. F. (2003). Juror decision-making in a mock Improving the predictive accuracy of Static-99 and Static-2002 with
sexually violent predator trial: Gender differences in the impact of older sex offenders: Revised age weights. Sexual Abuse: A Journal of
divergent types of expert testimony. Behavioral Sciences & the Law, Research and Treatment, 24(1), 64–101. https://doi.org/10.1177/
21(2), 215–237. https://doi.org/10.1002/bsl.529 1079063211409951
Guy, L. S., Edens, J. F., Anthony, C., & Douglas, K. S. (2005). Does Hemphill, J. F., Hare, R. D., & Wong, S. (1998). Psychopathy and
psychopathy predict institutional misconduct among adults? A meta- recidivism: A review. Legal and Criminological Psychology, 3(1),
analytic investigation. Journal of Consulting and Clinical Psychology, 139–170. https://doi.org/10.1111/j.2044-8333.1998.tb00355.x
73(6), 1056–1064. https://doi.org/10.1037/0022-006X.73.6.1056 Hemphill, J. F., Hart, S. D., & Hare, R. D. (1994). Psychopathy and
Hanson, R. K., & Morton-Bourgon, K. (2004). Predictors of sexual substance use. Journal of Personality Disorders, 8(3), 169–180.
recidivism: An updated meta-analysis (User Report 2004-02). Public https://doi.org/10.1521/pedi.1994.8.3.169
Safety and Emergency Preparedness Canada. Hildebrand, M., de Ruiter, C., & de Vogel, V. (2004). Psychopathy and
Hanson, R. K., & Morton-Bourgon, K. (2005). The characteristics of sexual deviance in treated rapists: Association with sexual and non-
persistent sexual offenders: A meta-analysis of recidivism studies. sexual recidivism. Sexual Abuse: A Journal of Research and
Journal of Consulting and Clinical Psychology, 73(6), 1154–1163. Treatment, 16(1), 1–24. https://doi.org/10.1177/107906320401600101
https://doi.org/10.1037/0022-006X.73.6.1154 Hogan, N., & Ennis, L. (2010). Assessing risk for forensic psychiatric
Hanson, R. K., & Morton-Bourgon, K. E. (2009). The accuracy of inpatient violence: A meta-analysis. Open Access Journal of Forensic
recidivism risk assessments for sexual offenders: A meta-analysis of Psychology, 2, 137–147. https://psycnet.apa.org/doi/10.1037/lhb0000179
118 prediction studies. Psychological Assessment, 21(1), 1–21. https:// Hollerbach, P., Habermeyer, E., Nitschke, J., S€ unkel, Z., & Mokros, A.
doi.org/10.1037/a0014421 (2020). Construct validity of the German version of the Psychopathy
Hanson, R. K., Helmus, L., & Bourgon, G. (2007). The validity of risk Checklist-Revised. European Journal of Psychological Assessment,
assessments for intimate partner violence: A meta-analysis. 36(5), 805–816. https://doi.org/10.1027/1015-5759/a000566
Howard, R., Khalifa, N., Duggan, C., & Lumsden, J. (2012). Are risk assessment. Psychological Assessment, 24(2), 524–529. https://
patients deemed “dangerous and severely personality disordered” doi.org/10.1037/a0026015
different from other personality disordered patients detained in Murrie, D. C., Boccaccini, M. T., Turner, D. B., Meeks, M., Woods, C.,
forensic settings? Criminal Behaviour and Mental Health: CBMH, & Tussey, C. (2009). Rater (dis)agreement on risk assessment meas-
22(1), 65–78. https://doi.org/10.1002/cbm.827 ures in sexually violent predator proceedings: Evidence of adversar-
Ismail, G., & Looman, J. (2018). Field inter-rater reliability of the ial allegiance in forensic evaluation? Psychology, Public Policy, and
Psychopathy Checklist-Revised. International Journal of Offender Law, 15(1), 19–53. https://doi.org/10.1037/a0014897
Therapy and Comparative Criminology, 62(2), 468–481. https://doi. Neumann, C. S., Hare, R. D., & Pardini, D. A. (2015). Antisociality
org/10.1177/0306624X16652452 and the construct of psychopathy: Data from across the globe.
Jackson, R. L., & Hess, D. (2007). Evaluation for civil commitment of Journal of Personality, 83(6), 678–692. https://doi.org/10.1111/jopy.
sex offenders: A survey of experts. Sexual Abuse: A Journal of 12127
Research and Treatment, 19(4), 425–448. https://doi.org/10.1177/ Neumann, C. S., Kosson, D. S., & Salekin, R. T. (2007). Exploratory
107906320701900407 and confirmatory factor analysis of the psychopathy construct:
Koo, T. K., & Li, M. Y. (2016). A guideline of selecting and reporting Methodological and conceptual issues. In H. Herve & J. C. Yuille
intraclass correlation coefficients for reliability research. Journal of (Eds.), The psychopath: Theory, research, and practice (pp. 79–104).
Chiropractic Medicine, 15(2), 155–163. https://doi.org/10.1016/j.jcm. Lawrence Erlbaum Associates Publishers.
2016.02.012 Olver, M. E., & Riemer, E. K. (2021). High-psychopathy men with a
Kumho Tire Co. v. Carmichael, 526 U.S. 137 (1999). history of sexual offending have protective factors too: But are these
Landis, J. R., & Koch, G. G. (1977). The measurement of observer risk relevant and can they change in treatment? Journal of
agreement for categorical data. Biometrics, 33(1), 159–174. https:// Consulting and Clinical Psychology, 89(5), 406–420. https://doi.org/
doi.org/10.2307/2529310 10.1037/ccp0000638
Langton, C. M., Barbaree, H. E., Harkins, L., & Peacock, E. J. (2006). Olver, M. E., & Wong, S. C. P. (2006). Psychopathy, sexual deviance,
Sex offenders’ response to treatment and its association with recidiv- and recidivism among sex offenders. Sexual Abuse: A Journal of
ism as a function of psychopathy. Sexual Abuse: A Journal of Research and Treatment, 18(1), 65–82. https://doi.org/10.1177/
Research and Treatment, 18(1), 99–120. https://doi.org/10.1007/ 107906320601800105
s11194-006-9004-5 Olver, M. E., & Wong, S. C. P. (2011). Predictors of sex offender treat-
Leistico, A. M. R., Salekin, R. T., DeCoster, J., & Rogers, R. (2008). A ment dropout: Psychopathy, sex offender risk and responsivity
large-scale meta-analysis relating the Hare measures of psychopathy implications. Psychology, Crime & Law, 17(5), 457–471. https://doi.
to antisocial conduct. Law and Human Behavior, 32(1), 28–45. org/10.1080/10683160903318876
https://doi.org/10.1007/s10979-007-9096-6 Olver, M. E., & Wong, S. P. (2015). Short- and long-term recidivism
Lieberman, J. D., & Krauss, D. A. (2009). The effects of labeling, expert
prediction of the PCL-R and the effects of age: A 24-year follow-up.
testimony, and information processing mode on juror decisions in
Personality Disorders, 6(1), 97–105. https://doi.org/10.1037/
SVP civil commitment trials. Journal of Investigative Psychology and
per0000095
Offender Profiling, 6(1), 25–41. https://doi.org/10.1002/jip.91
Olver, M. E., Lewis, K., & Wong, S. C. P. (2013). Risk reduction treat-
Lilienfeld, S. O., Watts, A. L., Smith, S. F., Berg, J. M., & Latzman,
ment of high-risk psychopathic offenders: The relationship of psych-
R. D. (2015). Psychopathy deconstructed and reconstructed:
opathy and treatment change to violent recidivism. Personality
Identifying and assembling the personality building blocks of
Disorders, 4(2), 160–167. https://doi.org/10.1037/a0029769
Cleckley’s Chimera. Journal of Personality, 83(6), 593–610. https://
Olver, M. E., Neumann, C. S., Sewall, L. A., Lewis, K., Hare, R. D., &
doi.org/10.1111/jopy.12118
Wong, S. C. P. (2018). A comprehensive examination of the psycho-
Lloyd, C. D., Clark, H. J., & Forth, A. E. (2010). Psychopathy, expert
metric properties of the Hare Psychopathy Checklist-Revised in a
testimony and indeterminate sentences: Exploring the relationship
Canadian multisite sample of Indigenous and non-Indigenous
between Psychopathy Checklist-Revised testimony and trial outcome
in Canada. Legal and Criminological Psychology, 15(2), 323–339. offenders. Psychological Assessment, 30(6), 779–792. https://doi.org/
https://doi.org/10.1348/135532509X468432 10.1037/pas0000533
Loving, J. E. (2002). Treatment planning with the Psychopathy Olver, M. E., Stockdale, K. C., Neumann, C. S., Hare, R. D., Mokros,
Checklist-Revised (PCL-R). International Journal of Offender A., Baskin-Sommers, A., Brand, E., Folino, J., Gacono, C., Gray,
Therapy and Comparative Criminology, 46(3), 281–293. https://doi. N. S., Kiehl, K., Knight, R., Leon-Mayer, E., Logan, M., Meloy, J. R.,
org/10.1177/0306624X02463003 Roy, S., Salekin, R. T., Snowden, R., Thomson, N., … Yoon, D.
Matsushima, Y. (2016). The inter-rater reliability of the Psychopathy (2020). Reliability and validity of the Psychopathy Checklist-Revised
Checklist-Revised in practical field settings [Unpublished master’s in the assessment of risk for institutional violence: A cautionary
thesis]. Southern Illinois University Carbondale. note on DeMatteo et al. Psychology, Public Policy, and Law, 26(4),
McGraw, K. O., & Wong, S. P. (1996). Forming inferences about some 490–510. https://doi.org/10.1037/law0000256
intraclass correlation coefficients. Psychological Methods, 1(1), 30–46. Parent, G., Guay, J., & Knight, R. A. (2011). An assessment of long-
https://doi.org/10.1037/1082-989X.1.1.30 term risk of recidivism by adult sex offenders: One size does not fit
Messick, S. (1995). Validity of psychological assessment: Validation of all. Criminal Justice and Behavior, 38(2), 188–209. https://doi.org/10.
inferences from persons’ responses and performances as scientific 1177/0093854810388238
inquiry into score meaning. American Psychologist, 50(9), 741–749. Polaschek, D. L. L. (2014). Adult criminals with psychopathy: Common
https://doi.org/10.1037/0003-066X.50.9.741 beliefs about treatability and change have little empirical support.
Mokros, A., Vohs, K., & Habermeyer, E. (2014). Psychopathy and vio- Current Directions in Psychological Science, 23(4), 296–301. https://
lent reoffending in German-speaking countries: A meta-analysis. doi.org/10.1177/0963721414535211
European Journal of Psychological Assessment, 30(2), 117–129. Polaschek, D. L. L., & Wong, S. C P. (2020). Risk-reducing treatment
https://doi.org/10.1027/1015-5759/a000178 in high risk psychopathic and violent offenders. In J. S. Wormith,
Morrissey, C., Hogue, T. E., Mooney, P., Lindsay, W. R., Steptoe, L., L. A. Craig, & T. E. Hogue (Eds.) The Wiley handbook of what
Taylor, J., & Johnston, S. (2005). Applicability, reliability, and valid- works in violence risk management: Theory, research and practice
ity of the Psychopathy Checklist-Revised in offenders with intellec- (pp. 367–384). Wiley. https://doi.org/10.1002/9781119315933.ch18
tual disabilities: Some initial findings. International Journal of Poythress, N. G., Lilienfeld, S. O., Skeem, J. L., Douglas, K. S., Edens,
Forensic Mental Health, 4(2), 207–220. https://doi.org/10.1080/ J. F., Epstein, M., & Patrick, C. J. (2010). Using the PCL-R to help
14999013.2005.10471225 estimate the validity of two self-report measures of psychopathy
Murrie, D. C., Boccaccini, M. T., Caperton, J., & Rufino, K. (2012). with offenders. Assessment, 17(2), 206–219. https://doi.org/10.1177/
Field validity of the Psychopathy Checklist-Revised in sex offender 1073191109351715
Quinsey, V. L., Harris, G. T., Rice, M. E., & Cormier, C. A. (2006). Slobogin, C., Hafemeister, T. L., & Mossman, D. (2020). Law and the
Violent offenders: Appraising and managing risk. American mental health system: Civil and criminal aspects (7th ed.). West
Psychological Association. Publishing.
Rice, M. E., & Harris, G. T. (1997). Cross-validation and extension of Sullivan, E. A., Abramowitz, C. S., Lopez, M., & Kosson, D. S. (2006).
the violence risk appraisal guide for child molesters and rapists. Law Reliability and construct validity of the Psychopathy Checklist-
and Human Behavior, 21(2), 231–241. https://doi.org/10.1023/ Revised for Latino, European American, and African American male
A:1024882430242 inmates. Psychological Assessment, 18(4), 382–392. https://doi.org/10.
Rice, M. E., & Harris, G. T. (2005). Comparing effect sizes in follow- 1037/1040-3590.18.4.382
up studies: ROC Area, Cohen’s d, and r. Law and Human Behavior, United States v. Sablan, 555 F. Supp. 2d 1177 (D. Colo. 2006).
29(5), 615–620. https://doi.org/10.1007/s10979-005-6832-7 van der Put, C. E., Gubbels, J., & Assink, M. (2019). Predicting domes-
Rosenthal, R. (1990). How are we doing in soft psychology? American tic violence: A meta-analysis on the predictive validity of risk assess-
Psychologist, 45(6), 775–777. https://doi.org/10.1037/0003-066X.45.6. ment tools. Aggression and Violent Behavior, 47, 100–116. https://
775 doi.org/10.1016/j.avb.2019.03.008
Rufino, K. A., Boccaccini, M. T., Hawes, S. W., & Murrie, D. C. Viljoen, J. L., MacDougall, E. A. M., Gagnon, N. C., & Douglas, K. S.
(2012). When experts disagreed, who was correct? A comparison of (2010). Psychopathy evidence in legal proceedings involving adoles-
PCL-R scores from independent raters and opposing forensic cent offenders. Psychology, Public Policy, and Law, 16(3), 254–283.
experts. Law and Human Behavior, 36(6), 527–537. https://doi.org/ https://doi.org/10.1037/a0019649
10.1037/h0093988 Walsh, T., & Walsh, Z. (2006). The evidentiary introduction of
Salekin, R. T., Rogers, R., & Sewell, K. W. (1996). A review and meta- Psychopathy Checklist-Revised assessed psychopathy in U.S. courts:
analysis of the Psychopathy Checklist and Psychopathy Checklist- Extent and appropriateness. Law and Human Behavior, 30(4),
Revised: Predictive validity of dangerousness. Clinical Psychology: 493–507. https://doi.org/10.1007/s10979-006-9042-z
Science and Practice, 3(3), 203–215. https://doi.org/10.1111/j.1468- Walters, G. D. (2003a). Predicting criminal justice outcomes with the
2850.1996.tb00071.x Psychopathy Checklist and Lifestyle Criminality Screening Form: A
Sanchez-Meca, J., Marın-Martınez, F., & Chacon-Moscoso, S. (2003). meta-analytic comparison. Behavioral Sciences & the Law, 21(1),
Effect-Size Indices for Dichotomized Outcomes in Meta-Analysis. 89–102. https://doi.org/10.1002/bsl.519
Psychological Methods, 8(4), 448–467. https://doi.org/10.1037/1082- Walters, G. D. (2003b). Predicting institutional adjustment and recidiv-
989X.8.4.448
ism with the psychopathy checklist factor scores: A meta-analysis.
Sandvik, A. M., Hansen, A. L., Kristensen, M. V., Johnsen, B. H.,
Law and Human Behavior, 27(5), 541–558. https://doi.org/10.1023/
Logan, C., & Thornton, D. (2012). Assessment of psychopathy:
A:1025490207678 https://doi.org/10.1023/A:1025490207678
Inter-correlations between Psychopathy Checklist Revised,
Weinstein, J. B., & Berger, M. A. (2015). Weinstein’s evidence manual:
Comprehensive Assessment of Psychopathic Personality-Institutional
A guide to the Federal Rules of Evidence based on Weinstein’s
Rating Scale, and Self-Report of Psychopathy Scale-III. International
Federal Evidence (10th ed.). Matthew Bender & Company.
Journal of Forensic Mental Health, 11(4), 280–288. https://doi.org/10.
Wong, S. (1988). Is Hare’s Psychopathy Checklist reliable without the
1080/14999013.2012.746756
interview? Psychological Reports, 62(3), 931–934. https://doi.org/10.
Sewall, L. A., & Olver, M. E. (2019). Psychopathy and treatment out-
2466/pr0.1988.62.3.931
come: Results from a sexual violence reduction program. Personality
Wong, S. C. P. (2016). Treatment of violence prone individuals with
Disorders, 10(1), 59–69. https://doi.org/10.1037/per0000297
Simourd, D. J., & Hoge, R. D. (2000). Criminal psychopathy: A risk- psychopathic personality traits. In J. Livesley, G. Dimaggio, & J.
and-need perspective. Criminal Justice and Behavior, 27(2), 256–272. Clarkin (Eds.), Integrated treatment of personality disorder: A modu-
https://doi.org/10.1177/0093854800027002007 lar approach (pp. 345–376). Guildford.
Singh, J. P., Desmarais, S. L., Hurducas, C., Arbach-Lucioni, K., Wong, S. C. P., & Gordon, A. (2006). The validity and reliability of the
Condemarin, C., Dean, K., Doyle, M., Folino, J. O., Godoy-Cervera, Violence Risk Scale: A treatment-friendly violence risk assessment
V., Grann, M., Mei Yee Ho, R., Large, M. M., Nielsen, L. H., Pham, tool. Psychology, Public Policy, and Law, 12(3), 279–309. https://doi.
T. H., Rebocho, M. F., Reeves, K. A., Rettenberger, M., de Ruiter, org/10.1037/1076-8971.12.3.279
C., Seewald, K., & Otto, R. K. (2014). International perspectives on Wong, S. C. P., Gordon, A., Deqiang, G., Lewis, K., & Olver, M. E.
the practical application of violence risk assessment: A global survey (2012). The effectiveness of violence reduction treatment for psycho-
of 44 countries. International Journal of Forensic Mental Health, pathic offenders: Empirical evidence and a treatment model.
13(3), 193–206. https://doi.org/10.1080/14999013.2014.922141 International Journal of Forensic Mental Health, 11(4), 336–349.
Singh, J. P., Grann, M., & Fazel, S. (2011). A comparative study of vio- https://doi.org/10.1080/14999013.2012.476760 https://doi.org/10.
lence risk assessment tools: A systematic review and metaregression 1080/14999013.2012.746760
analysis of 68 studies involving 25,980 participants. Clinical Wong, S., & Gordon, A. (1999–2003). The Violence Risk Scale (VRS).
Psychology Review, 31(3), 499–513. https://doi.org/10.1016/j.cpr.2010. Regional Psychiatric Centre and University of Saskatchewan.
11.009 Yang, M., Wong, S. C. P., & Coid, J. (2010). The efficacy of violence
Skeem, J. L., Johansson, P., Andershed, H., Kerr, M., & Eno Louden, J. prediction: A meta-analytic comparison of nine risk assessment
(2007). Two subtypes of psychopathic violent offenders that parallel tools. Psychological Bulletin, 136(5), 740–767. https://doi.org/10.
primary and secondary variants. Journal of Abnormal Psychology, 1037/a0020473
116(2), 395–409. https://doi.org/10.1037/0021-843X.116.2.395 Zwets, A. J., Hornsveld, R. H. J., Neumann, C., Muris, P., & van Marle,
Skeem, J. L., Monahan, J., & Mulvey, E. P. (2002). Psychopathy, treat- H. J. C. (2015). The four-factor model of the Psychopathy Checklist-
ment involvement, and subsequent violence among civil psychiatric Revised: Validation in a Dutch forensic inpatient sample.
patients. Law and Human Behavior, 26(6), 577–603. https://doi.org/ International Journal of Law and Psychiatry, 39, 13–22. https://doi.
10.1023/A:1020993916404 https://doi.org/10.1023/A:1020993916404 org/10.1016/j.ijlp.2015.01.016

Use of The Psychopathy Checklist-Revised in Legal Contexts Validity, Reliability

Uploaded by

Copyright:

Available Formats

Use of The Psychopathy Checklist-Revised in Legal Contexts Validity, Reliability

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

Use of The Psychopathy Checklist-Revised in Legal Contexts Validity, Reliability

Uploaded by

Copyright:

Available Formats

Journal of Personality Assessment

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/hjpa20

Use of the Psychopathy Checklist-Revised in Legal

David DeMatteo & Mark E. Olver

To link to this article: https://doi.org/10.1080/00223891.2021.1955693

Published online: 06 Aug 2021.

Submit your article to this journal

Article views: 4003

View related articles

View Crossmark data

Citing articles: 2 View citing articles

Full Terms & Conditions of access and use can be found at

Use of the Psychopathy Checklist-Revised in Legal Contexts: Validity, Reliability,

ABSTRACT ARTICLE HISTORY

You might also like