Measures of psychopathy

Introduction

In 1980, Robert D Hare first published the Psychopathy Checklist: a set of 22 behaviours or inclinations that may indicate psychopathy.  The tool was primarily derived from his observations with male offenders and inmates in Vancouver as well as traits that Cleckley (1941) had delineated previously to characterise psychopathy.  In 1991, Hare then revised this instrument to generate the Psychopathy Checklist-Revised that comprises 20 items (Hare, 1991). 

To administer the checklist, a mental health professional—often a psychologist or psychiatrist—will determine whether each of the 20 behaviours or inclinations fully applies, somewhat applies, or does not apply to the target participant.  To answer these questions, the mental health professional will conduct interviews with the participant as well scrutinise case files or other information about the person.  The items appear on various websites and include

• glib and superficial charm—such as slick and smooth but often unconstrained by social conventions, like the need to switch between talking and listening,
• conning and manipulative—such as the inclination to cheat or defraud other people ruthlessly
• callous rather than empathic, such as contemptuous, inconsiderate, and tactless rather than warm
• grandiose self-worth—in which people inflate their capabilities and seem opinionated,
• need for stimulation and proneness to boredom—such as a tendency to pursue thrilling activities and to shift jobs frequently,
• shallow affect, such as limited range or intensity of feelings including a sense of coldness even if gregarious,
• impulsivity, such as reckless acts because of an inability to resist urges and plan carefully,
• irresponsible behaviour, including repeated failures to honour commitments, such as pay bills or arrive on time,
• impaired behavioural control, such as impatience and verbal abuse,
• criminal versatility, such as a diversity of criminal offences or pride in averting convictions,
• juvenile delinquency.

For each item, the mental health professional assigns a score of 0, 1, or 2, depending on whether the behaviour or inclination does not apply, somewhat applies, or fully applies to the target participant. Accordingly, each participant receives a score that ranges from 0 to 40.  Typically, participants whose score exceeds 30 are classified as psychopathic (e.g., DeMatteo et al., 2014; for a diverging approach, see Balsis et al., 2017).   When this criterion is applied to diagnose psychopathy

• about 1% of people who reside in Western nations fulfill this definition of psychopathy (Coid et al., 2009; Neumann & Hare, 2008),
• about 20% of prisoners in these nations fulfill this definition (Hare, 2003).

Two factors

As preliminary studies indicated (e.g., Harpur et al., 1988; Harpur et al., 1989), these items can be divided into two clusters or factors.  The first factor primarily revolves around selfish interpersonal behaviour and entails, for example, superficial charm, conning and manipulative behaviour, callous rather than empathic, and shallow affect.  These tendencies seem to correspond more to the traditional characterisation of psychopathy, sometimes called primary psychopathy.

The second factor primarily revolves around an unstable, criminal, and antisocial lifestyle.  To illustrate, this factor entails impulsivity, irresponsible behaviour, impaired behavioural control, and juvenile delinquency.  In contrast to the first factor, these tendencies often coincide with criminal behaviour, family background, and antisocial personality disorder.

Typically, the two factors, although positively associated with each other (Harpur et al., 1989), differ in their relationships with measures of wellbeing or thriving.  For example, the first factor is not statistically associated with suicide risk, whereas the second factor is positively associated with suicide risk (Verona et al., 2001).

The two factors also coincide with distinct criminal acts.  For example, as Hart and Dempster (1997) revealed, the first factor tends to predict criminal acts that are planned carefully.  In contrast, the second factor is more likely to predict criminal acts that are impulsive and spontaneous, often emanating from strong emotions, such as anger or anxiety.el also excluded the items that revolve around criminal behaviour, partly because the instrument is designed to predict, rather than describe, such acts. 

Three factors

Yet, as subsequent research has revealed, the items might correspond to more than two factors.  According to some researchers (e.g. Cooke & Michie, 2001), previous attempts to apply confirmatory factor analysis to substantiate these factors have been unsuccessful.  For instance, in a sample of 376 participants in Australia, the model fit indices were inadequate (Darke et al., 1998).  Similarly, in another study, the comparative fit index was only .83—typically suggestive of inadequate fit although the authors deemed this fit as moderate (Brandt et al., 1997).  

In 2001, Cooke and Michie, two academics employed at the Glasgow Caledonian University, conducted a study that derived three factors from data collected in Scotland and North America. The three factors can be interpreted as

• an arrogant and deceitful interpersonal style, comprising items such as glibness and superficial charm, a grandiose sense of self-worth, pathological lying, as well as conning and manipulative behaviour,
• deficient affective experience, comprising items such as shallow affect, limited empathy, limited remorse or guilt, and a failure to accept responsibility,
• impulsive and irresponsible behaviour, comprising items such as need for stimulation, impulsivity, irresponsible behaviour, and a parasitic lifestyle. 

In essence, this solution divides the first factor that previous researchers had identified into two factors, sometimes called 1A and 1B.  However, this model also excluded the items that revolve around criminal behaviour, partly because the instrument is designed to predict, rather than describe, such acts. 

Four factors

Later studies showed that 18 of the 20 items in the Psychopathy Checklist-Revised can be reduced to four distinct factors (for a summary, see Hare & Neumann 2008; for similar factors, see Paulhus et al., 2016). These factors or dimensions are called

• interpersonal manipulation—comprising glib and superficial charm, grandiose self-worth, pathological lying, as well as conning and manipulative.  
• callous affect— comprising limited remorse, shallow affect, callous rather than empathic, and failure to accept responsibility,
• erratic lifestyle—comprising need for stimulation, impulsivity, irresponsible behaviour, a parasitic orientation, and the absence of realistic goals,
• antisocial behaviour—comprising impaired behavioural control, early behaviour problems, juvenile delinquency, revocation of conditional release, and criminal versatility.

Interpersonal manipulation and callous affect both coincide with a shared factor, comparable to the first factor that Robert Hare had originally delineated.  Similarly, erratic lifestyle and antisocial behaviour coincide with another shared factor, comparable to the second factor that Robert Hare had originally delineated. All four dimensions also coincide with one super-ordinate factor, representing psychopathy overall (Neumann et al., 2007). Many studies have corroborated these four factors in diverse samples, such as

• adolescent offenders (Jones et al., 2006; Salekin et al., 2006)—although some of these studies have also validated a model that comprises three factors as well (Jones et al., 2006; Salekin et al., 2006),
• inpatient psychiatric patients disorders (Hill et al., 2004), even after controlling ethnicity, gender, and intelligence (Vitacco et al., 2005),
• both European Americans and African Americans (Jackson et al., 2007)

Typically, to delineate these four dimensions, researchers usually refer to the presence of undesirable qualities or inclinations.  In contrast, Curtis et al. (2025) identified the absence of desirable qualities or inclinations that correspond to each dimension.  Specifically, Curtis et al. proposed that

• interpersonal manipulation can partly be defined as a deficit in honesty and humility—because deception is the converse of honesty and grandiosity is the converse of humility,
• callous affect can partly be defined as a deficit in empathy,
• erratic lifestyle can partly be defined as a deficit in conscientiousness,
• antisocial behaviour in general can partly be defined as a deficit in regulatory governance—or the capacity to comply with social, cultural, and legal rules.

Concerns about the fourth factor

The fourth factor of this model—called antisocial behaviour and comprising items like juvenile delinquency and criminal versatility—has attracted considerable debate and criticism. As several commentators in this field have underscored (e.g., Skeem & Cooke, 2010), these items, in essence, revolve around criminal behaviour.  The problem is that

• to complete the tool, practitioners need to ascertain whether the target participant has perpetrated criminal behaviour,
• if the target participant has perpetrated criminal behaviour and thus scores high on the Psychopathy Checklist-Revised, the practitioner then predicts the individual has perpetrated criminal behaviour—a tautological conclusion (Almas & Lordos, 2025).

Concerns around the scoring: Limited weighting of affective features

Scholars have raised a concern about the scoring of the Psychopathy Checklist-Revised—and, specifically, the equal weighting that is applied to each item.  The problem is that, according to many forensic practitioners, some of the items are more representative or prototypical of psychopathic individuals than other items.  Yet, because all items are weighted equally, some participants might be diagnosed as psychopathic but may not demonstrate any of the core features.

To illustrate, in one study, published in 2020 by two academics, Verschuere and Te Kaat, at the University of Amsterdam, 57 forensic mental health practitioners evaluated the degree to which each item of the Psychopathy Checklist-Revised is relevant to the diagnosis of psychopathy.  As this study revealed, the items that revolve around deficient affective experiences—such as limited empathy, remorse, or guilt—and, to a lesser extent interpersonal problems—such as conning and manipulative—were deemed as core to this diagnosis (for similar conclusions, but derived from a network approach, see McCuish et al., 2019).  Yet, in practice, individuals could be diagnosed with psychopathy without demonstrating these core features.  Arguably, to overcome this problem, the scoring could be adjusted:

• Perhaps, regardless of the score, individuals should not be classified as psychopathic unless they demonstrate some of these affective or interpersonal concerns.
• Alternatively, these affective or interpersonal dimensions could be assigned a higher weight. 

Capacity to predict violence

The degree to which the Psychopathy Checklist-Revised predicts crimes has been studied carefully.  Although sometimes useful, other tools may predict specific crimes more effectively. For example, as a systematic review and meta-regression revealed (Singh et al., 2011), several tools predict violent crimes more effectively than does the Psychopathy Checklist-Revised.  Specifically, Singh et al. (2011) evaluated several instruments in addition to the Psychopathy Checklist-Revised, including

• the Historical, Clinical, Risk Management-20 (Webster & Eaves, 1995),
• the Level of Service Inventory-Revised (Andrews & Bonta, 1995),
• the Static-99 (Hanson & Thornton, 1999),
• the Structured Assessment of Violence Risk in Youth (Borum et al., 2003), and
• the Violence Risk Appraisal Guide (Quinsey et al., 2000).

The review analysed data that were derived from 68 studies, almost 26 000 participants, and 13 nations.   Overall, the Structured Assessment of Violence Risk in Youth, designed to predict violence in adolescents, was generally the most useful.  In contrast, the Psychopathy Checklist-Revised and Level of Service Inventory-Revised generated the least accurate predictions—perhaps because these tools were designed to predict a diverse range of offences across many ages and settings.  Tools that are designed to predict a particular offence in a subset of the population may thus be more effective.

Introduction

Patrick (2010) constructed and evaluated the triarchic psychopathy measure partly to circumvent some limitations of the Psychopathy Checklist-Revised.  For example,

• whereas the Psychopathy Checklist-Revised can demand extensive time and investigation to complete (Copestake et al., 2011), the triarchic psychopathy measure is shorter and simpler,
• whereas the Psychopathy Checklist-Revised primarily revolves around criminal acts (for a discussion, see Skeem & Cooke, 2010), the triarchic psychopathy measure assesses the underlying attributes.

The triarchic psychopathy can be accessed from this website. This measure was derived from a taxonomy, formulated by Patrick et al. (2009), that characterises the developmental origins of three key dimensions of psychopathy: disinhibition, boldness, and meanness. According to this account

• disinhibition revolves around impairments in the capacity of individuals to regulate their impulses or temptations,
• boldness combines a sense of dominance in social interactions as well as resilience and risky behaviour in other settings,
• meanness relates to a tendency of individuals to fulfill personal interests aggressively with negligible regard for the needs of other people.

The measure

As Patrick et al. (2009) argue, disinhibition and boldness seem to emanate primarily from a difficult temperament—that is, an inherent nature or personality that entails significant negative affect and irritability, coupled with high levels of activity and impairments in the capacity to adapt effectively to changes in the environment.  These tendencies could be ascribed to impairments in the frontal-cortical circuits that regulate behaviour as well as limited activity in the defensive motivation systems of the brain (Patrick et al., 2012).  In contrast, boldness and meanness seem to emanate from low dispositional fear—such as a tendency to approach, rather than avoid, risky situations—as well as limited empathy.  The limited empathy could partly be ascribed to dysfunction in oxytocin, resinessin, and other endogenous neuromodulators (Patrick et al., 2012). To measure these dimensions, Patrick (2010) constructed a scale that comprises

• 19 items to measure boldness, such as “I am afraid of far fewer things than most people”, “I can get over things that would traumatize others”, and “It worries me to go into an unfamiliar situation without knowing all the details” (reverse scored),
• 20 items to measure disinhibition, such as “I have lost a friend because of irresponsible things I’ve done”, “I have a hard time waiting patiently for things I want”, and “I get in trouble for not considering the consequences of my actions”,
• 19 items to assess meanness, such as “I don’t see any point in worrying if what I do hurts someone else”, “I taunt people just to stir things up”, and “How other people feel is important to me” (reverse scored).

Evidence of reliability and validity

To establish the validity of this instrument, Patrick (2010) administered these items to a sample of 94 college students, together with several other measures including

• the Psychopathic Personality Inventory (Lilienfeld & Widows, 2005), designed to assess all facets of psychopathy in the general population
• the Hare Self-Report Psychopathy Scale, Version 2 (Williams et al., 2007), similar to the Psychopathy Checklist, except that individuals rate themselves on the various traits,
• the Youth Psychopathic Traits Inventory
• the Levenson Self-Report Psychopathy Scale, and
• the NEO antagonism facet.

As Patrick (2010) revealed, responses to the boldness, meanness, and disinhibition subscales were all moderately to highly correlated with scores on the Psychopathic Personality Inventory, the Hare Self-Report Psychopathy Scale, and the Youth Psychopathic Traits Inventory.  The Levenson Self-Report Psychopathy Scale and NEO antagonism subscale were highly associated with meanness and disinhibition but not boldness.

Studies have also explored the internal consistency and factor structure.  For example, Sellbom and Phillips (2013) administered this instrument to a sample of 209 female prisoners as well as 627 undergraduate students. The prisoners generated alpha internal consistency values of .89, .90, and .89 for boldness, meaning, and disinhibition respectively.  The undergraduate students generated alpha internal consistency values of .82, .88, and .84 for boldness, meaning, and disinhibition respectively. 

Nevertheless, in one study, in which over 1000 members of the public completed this instrument, confirmatory factor analysis did not substantiate the three factors: CFI = .76 (Roy et al 2021). Instead, this study uncovered seven factors in which

• positive self-image, leadership, and stress immunity were derived from the items that purportedly measure boldness,
• callousness and enjoy hurting were derived from the items that purportedly assess meanness, and
• trait impulsivity and overt antisociality were derived from the items that purportedly measure disinhibition.

Studies have also explored the association between these sub-scales and other relevant personality traits (for reviews, see Evans & Tully, 2016; Patrick & Drislane, 2015).  For example, Sellbom and Phillips (2013) revealed that

• boldness tends to coincide with narcissism, adventure seeking, and limited behaviour inhibition,
• meanness tends to coincide with limited empathy and Machiavellianism, and
• disinhibition tends to coincide with impulsivity and the pursuit of fun.

This pattern diverges, at least marginally, from the hypotheses. For example, meanness was more related to sensation seeking than anticipated. Risk in Youth, designed to predict violence in adolescents, was generally the most useful.  In contrast, the Psychopathy Checklist-Revised and Level of Service Inventory-Revised generated the least accurate predictions—perhaps because these tools were designed to predict a diverse range of offences across many ages and settings.  Tools that are designed to predict a particular offence in a subset of the population may thus be more effective.

Short version

Researchers have developed many instruments to measure psychopathy, such as the PCL-R.  However, when researchers want to assess psychopathy in the workplace, most of these instruments may be unsuitable for several reasons.  Specifically

• in some instances, only a clinician or specialist may complete the instrument,
• in other instances, individuals must complete the instrument to evaluate themselves—and these evaluations may be biased or inaccurate,
• few instrument comprise items that are specifically designed to assess behaviours in the workplace.

To overcome these limitations, Babiak and Hare (2012) developed the B-Scan 360. This measure comprises two distinct features:

• first, all the items refer to behaviours or attitudes that HR personnel or organisational psychologists deemed as concerning in the workplace—such as “threatens coworkers” or “asks harsh questions”,
• second, usually other people, such as supervisors, peers, or subordinates, evaluate the participants.

To explore and to characterise the factors or subscales of the B-Scan 360, Mathieu et al. (2013) conducted two studies.  To conduct the first study, 340 working adults, recruited from Amazon Mechical Turk, were invited to rate the personality of their supervisor or boss on 113 attributes, such as “Shows no regret for making decisions that harm the company, shareholders, or employees”.  To extract the factors, the researchers applied parallel analysis (Horn, 1965) and the minimal average partial (Zwick & Velicer, 1986) criteria to both the Pearson and polychoric correlation matrices (for a justification, see Cho et al., 2009)—primarily because distributions of psychopathy scores do not usually conform to a normal distribution.  Next, to complete the exploratory factor analysis, the researchers applied the weighted least squares estimation procedure.

Although the analysis uncovered six factors, two of these factors revolved more around performance and potential rather than psychopathy. When items that loaded on these two factors or items that loaded on no factors were removed, the remaining 38 items were subjected to another factor analysis.  This analysis generated four factors.  For each factor, the researchers then decided to retain only the five items that produced the highest loadings.  These factors roughly overlap with the four-factor model of the PCL-R and comprise

• manipulative and unethical—such as glib, uses charm, claims expertise, ingratiates himself or herself, and rationalises inappropriate behaviour,
• callous and insensitive—such as rarely shows emotions, insensitive, cold inside, remorseless, and no empathy,
• unreliable and unfocussed—such as disloyal, absence of planning, unfocussed, impatient, and unreliable,
• intimidating and aggressive—such as intimidating, angry, asks harsh questions, threatens co-workers, and dramatic.

The second study, in which the participants were 806 working adults, generated data that confirmed these factors, in which the Tucker-Lewis Index was .93 and the standardized root mean square residual was .07. 

Long version

Besides this short version of the B-Scan 360, researchers have also developed and validated a long version (Mathieu & Babiak, 2016).  The long version comprises 15 facets, and four items assess each facet.  Each facet corresponds to one of the four overarching factors.  Specifically

• insincere, arrogant, untrustworthy, and manipulative corresponds to manipulative and unethical,
• remorseless, shallow, insensitive, and blaming corresponds to callous and insensitive,
• impatient, selfish, unfocussed, erratic, and unreliable corresponds to unreliable and unfocussed,
• dramatic and bullying corresponds to intimidating and aggressive.

Babiak and Mathieu (2025) conducted two studies to assess whether this long version could be applied to measure other people in the workplace, such as managers.  For the first study, over 500 working adults completed the instrument to assess their immediate supervisor.  Confirmatory factor analysis tentatively validated the 15 facets and four underlying factors: the Tucker-Lewis Index was .85 and the standardized root mean square residual was .06. 

For the second study, 340 working adults completed this instrument as well as a measure of the five personality traits—again to assess their supervisor and not themselves.  As the findings revealed, all four factors were inversely related to agreeableness, conscientiousness, and emotional stability.