Clarifying the Content Coverage of Differing Psychopathy Inventories through Reference to the Triarchic Psychopathy Measure

Abstract

The Triarchic Model of psychopathy (Patrick, Fowles, and Krueger, 2009) was formulated as an integrative framework for reconciling differing conceptions of psychopathy. The model characterizes psychopathy in terms of three distinguishable phenotypic components: boldness, meanness, and disinhibition. Data from a large mixed-gender undergraduate sample (N = 618) were used to examine relations of several of the best-known measures for assessing psychopathic traits with scores on the Triarchic Psychopathy Measure (TriPM), an inventory developed to operationalize the Triarchic Model through separate facet scales. Analyses revealed that established inventories of psychopathy index components of the model as indexed by the TriPM to varying degrees. While each inventory provided effective coverage of meanness and disinhibition components, instruments differed in their representation of boldness. Current results demonstrate the heuristic value of the Triarchic Model for delineating commonalities and differences among alternative measures of psychopathy, and provide support for the utility of the Triarchic Model as a framework for reconciling alternative conceptions of psychopathy.

Introduction

Psychopathy is generally conceptualized as entailing persistent behavioral deviance in conjunction with deficient affect and impaired social relatedness. However, the precise definition and boundaries of the disorder, and the most appropriate methods for assessing it, remain topics of ongoing debate. Patrick, Fowles, and Krueger (2009) formulated a Triarchic Model in an effort to clarify and reconcile various alternative conceptions of psychopathy and provide a framework for understanding commonalities and differences among existing assessment methods and integrating empirical findings. As an initial approach to operationalizing the distinct phenotypic constructs described in the Triarchic Model, Patrick (2010) created the Triarchic Psychopathy Measure (TriPM), a self-report-based inventory containing 58-items organized into three facet scales: Boldness, Meanness, and Disinhibition. The present study sought to clarify similarities and differences among established inventories of psychopathy through reference to their coverage of content domains assessed by the TriPM, and in the process delineate more clearly distinct dimensional facets of psychopathy indexed by various assessment instruments

Alternative Conceptions of Psychopathy in the Literature

Psychopathy investigators have long sought to identify the constellation of traits central to this multifaceted disorder. Hervey Cleckley’s Mask of Sanity (1976), for example, has served as an essential foundation for most modern conceptions of psychopathy. According to Cleckley, psychopathic individuals are characterized by a psychologically healthy outward appearance, which masks a deep-rooted emotional disturbance. This underlying disturbance gives rise to observable maladjustment in the form of recklessness, irresponsibility, egocentricity, and shallow/exploitative interpersonal relations. However, adaptive manifestations of this underlying affective pathology (including ostensible charm, absence of anxious-depressive or psychotic symptoms, and disinclination toward suicide) operate to conceal these more maladaptive tendencies. By contrast, other writers of Cleckley’s time placed greater emphasis on malevolence and cruelty in their descriptions of psychopathy. For example, McCord and McCord (1964) identified traits of “lovelessness” and “guiltlessness” as the core defining features of the disorder and portrayed psychopathic individuals as cold, vicious, and exploitative. Other prominent researchers posited fearless temperament as the substrate for psychopathic tendencies (Lykken, 1957). Notably, each of these historic perspectives defines psychopathy in largely characterological terms, emphasizing dimensions of maladaptive personality in preference or in addition to overtly criminal or otherwise antisocial behaviors.

At present, the most widely-used instrument for assessing the affective, interpersonal, and impulsive-antisocial traits described in historic conceptualizations of psychopathy in correctional and forensic settings is the interview-based Psychopathy Checklist-Revised (PCL-R; Hare, 1980, 2003). In contrast to Cleckley, the PCL-R does not include items directly reflecting positive adjustment features (Patrick, 2006). As the PCL-R is often labor-intensive to administer, self-report inventories have also been developed for assessing psychopathy in adults and younger samples. Whereas the PCL-R was developed for use in forensic settings, studies employing various psychopathy self-report measures have largely focused on non-correctional, community samples. Some of these instruments are modeled after the PCL-R, including the Self-Report Psychopathy Scale-version III (SRP-III; Pauhlus, Hemphill, & Hare, 2009), Youth Psychopathic Traits Inventory (YPI; Andershed, Kerr, Stattin, & Levander, 2002), Antisocial Process Screening Device (APSD; Frick & Hare, 2001), and Child Psychopathy Scale (CPS; Lynam, 1997). Others, such as the Psychopathic Personality Inventory (PPI; Lilienfeld & Andrews, 1996; Lilienfeld & Widows, 2005), were developed using a theory-neutral approach informed by various conceptualizations of psychopathy, including Cleckley’s original conception; as such, the PPI incorporates items directly indicative of positive adjustment. Other measures beyond these embody somewhat different conceptions (e.g., Levenson, Kiehl, & Fitzpatrick, 1995 Self-Report Psychopathy Scale [LSRP], designed to index “primary” and “secondary” variants) or focus on selected aspects of psychopathy (e.g., Inventory of Callous-Unemotional Traits [ICU]; Frick, 2004).

Importantly, existing psychopathy self-report inventories approach the assessment of psychopathy from a distinctly dimensional perspective and measure pathological features in terms of lower-order traits rather than discrete symptoms indicative of a categorical diagnostic entity. Empirically, structural analyses of various psychopathy measures have failed to consistently support a taxonic solution, particularly with regard to the affective and interpersonal features of the disorder, suggesting that psychopathy is best represented as a continuum (Cale & Lilienfeld, 2002; Marcus, John, & Edens, 2004). This approach is aligned with the dominant perspective on personality disorders (PDs) more broadly, which holds that PDs are better conceived of as configurations of dimensional constructs (i.e., extreme variants of normal personality traits) than as discrete categories, or taxa (Clark, 2007; Frances & Widiger, 2012; Livesly & Jang, 2000; Trull & Durrett, 2005). Consistent with this, dimensional approaches to assessing psychopathy capture the degree to which an individual displays prototypic components of the disorder, rather than categorizing individuals as psychopathic or non-psychopathic.¹ Variability in these traits can be observed across the entire range of severity of the underlying dimensions, allowing for research on the psychopathy continua to be extended to non-clinical community samples, including undergraduates (Lilienfeld, 1998). Nonetheless, disagreement persists regarding the particular dimensions that psychopathy encompasses (Miller & Lynam, 2012; Lilienfeld et al., 2012).

Triarchic Model and Measure

The Triarchic Model of Psychopathy (Patrick et al., 2009) was proposed as a dimensional framework for reconciling alternative theoretic conceptions and integrating findings across differing assessment instruments. According to this model, the essential phenotypic components of psychopathy are disinhibition (i.e., general propensity toward externalizing problems, entailing traits such as impulsivity, irresponsibility, and hostility; Krueger, Markon, Patrick, Benning, & Kramer, 2007), meanness (i.e., the “callous-aggression” subdomain of the externalizing spectrum [Krueger et al., 2007; Venables & Patrick, 2012], related to traits such as manipulativeness, lack of empathic concern, and cruelty), and boldness (i.e., the adaptive component of psychopathy, entailing traits of dominance, emotional stability, and venturesomeness). Whereas disinhibition is theorized to reflect dysfunction in anterior brain systems including the prefrontal cortex and anterior cingulate cortex, which serve to inhibit behaviors and regulate affect (Patrick, 2008), boldness is thought to reflect the phenotypic expression of fearless temperament, associated with deficient sensitivity of affective processing systems of the brain, including the amygdala (Patrick & Bernat, 2009). Along with some ingredient of fearlessness (Frick & Marsee, 2006), dispositional meanness is theorized to arise from constitutional and environmental influences that contribute to impaired capacity for affiliation and nurturance (Patrick, Drislane, & Strickland, 2012a).

The Triarchic Psychopathy Measure (TriPM; Patrick, 2010) was developed to operationalize the three facets of the Triarchic Model as distinct dispositional variables, in a time-efficient manner. The constituent subscales of the TriPM were developed with reference to structural models of externalizing psychopathology and fear/fearlessness. Specifically, the Disinhibition and Meanness subscales of the TriPM index separable disinhibitory and callous-aggressive dimensions from a model of impulse-related problems delineated by the Externalizing Spectrum Inventory (ESI; Krueger et al., 2007; Venables & Patrick, 2012). The items of the Disinhibition scale are from subscales of the ESI that load exclusively on the broad externalizing factor (or “disinhibitory proneness”; Venables & Patrick, 2012), whereas items of the Meanness scale are from ESI scales that show appreciable loadings on the callous-aggression factor of the ESI. The Boldness scale, noted in the manual for the TriPM (Patrick, 2010), consists of items developed as indicators of 9 distinct content areas, representing fearlessness as expressed in interpersonal (i.e., persuasiveness, social assurance, dominance), affective-experiential (resiliency, self-assurance, optimism), and behavioral venturesomeness domains (courage, intrepidness, and tolerance for uncertainty). In turn, these domains reflect distinct thematic facets of fear/fearlessness identified through structural modeling of measures of this type, including the three PPI-Fearless Dominance (PPI-FD; Benning et al., 2003) subscales along with several other scale indicators from other inventories (Kramer et al., 2012). As such, TriPM Boldness and PPI-Fearless Dominance can be conceived of as operationalizations of a similar construct; however, items of the TriPM Boldness scale were generated separately from the PPI and are not direct counterparts of items in the PPI.

Encouraging validation results for the TriPM have been reported. Sellbom and Phillips (2013) found TriPM Boldness, Meanness, and Disinhibition to be related in distinct and expected ways to psychopathy-relevant criterion variables in undergraduate and prisoner samples (e.g., Boldness with thrill-seeking, narcissism, and low anxiousness; Meanness with Machiavellianism and lack of empathy; and Disinhibition with measures of impulsivity and pleasure-seeking). Marion et al. (2013) factor analyzed the subscales of the TriPM in conjunction with subscales of the PPI and LSRP and found evidence for three factors reflecting the constructs of the Triarchic Model, which were anchored most strongly by the TriPM scales. In other work, Stanley et al. (2013) reported patterns of relations for the TriPM scales with normal-range personality variables that converged with and served to clarify findings from prior research on the personality correlates of psychopathy inventories (Benning, Patrick, Blonigen, Hicks, & Iacono, 2005). In particular, extroversion was uniquely related to Boldness, neuroticism was related in opposing directions to Boldness (−) and Disinhibition (+), agreeableness was related most strongly (in a negative direction) to Meanness, and conscientiousness was associated negatively with both Meanness and Disinhibition, but somewhat positively with Boldness. Additionally, Stanley et al. (2013) reported incremental validity for the TriPM in predicting personality traits of known relevance to psychopathy over and above a comparable-length short form of the PPI.

Hypotheses

The current study sought to clarify conceptions of psychopathy embodied in differing assessment inventories by assessing the extent to which they capture the phenotypic dimensions described in Triarchic Model of psychopathy (Patrick et al., 2009), using the TriPM as an operationalization of these constructs. Data for the TriPM and other psychopathy inventories were collected along with scores on facets of Agreeableness from the NEO Personality Inventory-Revised (NEO-PI-R; Costa & McCrae, 1992) and primary trait scales of the Multidimensional Personality Questionnaire (MPQ; Tellegen, 2011) from a large, mixed-gender sample of undergraduates.

Hypotheses for the normal-range personality variables, which were collected to further validate the TriPM with respect to variables in this domain (Stanley et al., 2013), were as follows:

• Hypothesis 1

  Among the Big Five dimensions, agreeableness reversed (i.e., antagonism) is considered especially central to psychopathy (Lynam & Derefinko, 2006), and Stanley et al., (2013) found antagonism to be most related to the meanness subscale of the TriPM. Accordingly, we predicted that scores on NEO-PI-R Antagonism and its facets would correlate most strongly with TriPM Meanness.

• Hypothesis 2

  Based on prior empirical work demonstrating associations between MPQ primary traits and distinctive components of psychopathy (Benning et al., 2003; Verona, Patrick, & Joiner, 2001), we predicted differing patterns of relations for the three facet scales of the TriPM with trait scales of the MPQ: (1) TriPM Boldness (akin to PPI-FD) would be associated with MPQ Social Potency, Stress Reaction (−), and Harm Avoidance (−); (2) TriPM Meanness would be associated with MPQ Aggression and Social Closeness (−); and (3) TriPM Disinhibition would be associated with MPQ Alienation, Stress Reaction, and Control (−). Evidence for associations of particular facets of psychopathy with the Traditionalism scale of the MPQ has been rather modest in prior published work (Benning et al., 2003; Verona et al., 2001); as such, we do not propose a priori hypotheses regarding which subscale(s) of the TriPM might relate most strongly to scores on MPQ Traditionalism.

Hypotheses for the various psychopathy inventories were as follows:

• Hypothesis 3

  As putative indices of a common diagnostic construct, we hypothesized that overall scores on the TriPM would correlate significantly with total scores on each of the adult and child psychopathy measures.

• Hypothesis 4

  The TriPM Disinhibition and Meanness scales were developed to index the general disinhibition and callous-aggression factors, respectively, of the Externalizing Spectrum Inventory (ESI; Krueger et al., 2007). Since psychopathy is conceived of as an externalizing disorder (Patrick, Hicks, Krueger, & Lang, 2005), we predicted that all of the omnibus adult and juvenile psychopathy measures would evidence significant correlations with both TriPM Disinhibition and Meanness.

• Hypothesis 5

  The Inventory of Callous-Unemotional Traits (ICU; Frick, 2004) was designed to index one distinctive domain of juvenile antisocial symptoms—that of callous-unemotional traits, the counterpart to adult constructs of callous-aggression, antagonism, and meanness (Frick & Ellis, 1999; Patrick et al., 2012a). As such, we predicted that scores on the ICU would be related to TriPM Meanness in particular. Additionally, in view of recent evidence (Berg et al., 2013) that scores on the ICU correlate more positively with indices of psychological distress than expected based on theory; we hypothesized that scores on the ICU might also show a secondary relationship with TriPM Disinhibition.

• Hypothesis 6

  Self-report inventories vary in the extent to which they index the adaptive features of psychopathy described by Cleckley. We hypothesized that the PPI, which explicitly indexes features of this type in its Fearless Dominance subscales, would demonstrate the strongest relationship with TriPM Boldness, as the subscales of the PPI-FD factor (and the broader fear/fearlessness factor on which they load; Kramer et al., 2012) served as referents for development of this TriPM facet scale

• Hypothesis 7

  We predicted that questionnaires modeled after the PCL-R (SRP-III, YPI, CPS, APSD) would also demonstrate associations with TriPM Boldness, but to a lesser degree--and primarily with subscales indexing the interpersonal features of psychopathy (Benning et al., 2005; Hall, Benning, & Patrick, 2004). We predicted that TriPM Meanness would show preferential relations with affective subscales of these inventories, and that TriPM Disinhibition would relate preferentially to subscales indexing impulsive-antisocial tendencies (Drislane, Patrick, Hall, Poythress, & Lilienfeld, 2012).

• Hypothesis 8

  We predicted that TriPM Boldness would correlate strongly and preferentially with the fearless-dominance factor of the PPI and its constituent subscales, Social Potency, Stress Immunity, and Fearlessness (Sellbom & Phillips, 2013). Given that TriPM Meanness and Disinhibition each contain content related to sensation-seeking (i.e., items indicative of excitement seeking and boredom proneness, respectively), we predicted that these scales would (along with TriPM Boldness) exhibit positive relations with PPI Fearlessness.

By contrast, we predicted a negative association for TriPM Disinhibition with PPI Stress Immunity, in view of evidence that externalizing proneness is associated with heightened negative affectivity (Blonigen et al., 2010; Krueger, Caspi, Moffitt, Silva, & McGee, 1996; Sher & Trull, 1994). Positive relations were predicted for TriPM Disinhibition with the impulsive-antisociality (IA) factor of the PPI and its constituent subscales (Machiavellian Egocentricity, Impulsive/Rebellious Nonconformity, Alienation, Carefree Nonplanfulness), given evidence that PPI-IA indexes externalizing proneness (Benning et al., 2005; Blonigen, Hicks, Krueger, Patrick, & Iacono, 2005; Patrick, Edens, Poythress, Lilienfeld, & Benning, 2006).

We further hypothesized that scores on PPI Coldheartedness would be strongly and uniquely associated with TriPM Meanness. An additional more tentative hypothesis (per Sellbom & Phillips, 2013), was that scores on TriPM Meanness would be related somewhat to scores on PPI-IA (i.e., to a weaker degree than TriPM Disinhibition), primarily through association with its Machiavellian Egocentricity facet scale (which contains items indicative of callous-aggressiveness).

• Hypothesis 9

  The item content of the LSRP appears consistent with aggressive, loveless conceptions of psychopathy (McCord & McCord, 1964) and contains limited representation of adaptive features (Cleckley, 1976; Lykken, 1995). As such, we predicted that overall LSRP scores would be strongly related to TriPM Meanness and Disinhibition, and unrelated to TriPM Boldness. We further hypothesized that the LSRP’s Primary subscale would be preferentially related to TriPM Meanness, whereas its Secondary subscale would be preferentially related to TriPM Disinhibition.

Method

Participants

Analyses were performed on data for 618 undergraduate psychology students recruited from a large public university in the Southeastern region of the United States (M age = 18.8, SD = 1.65; 56.2% female; 76.2% Caucasian, 8.9% African American, 1.5% Asian, 0.3% Native American, 2.6% Mixed or Other Race, 14.2% Hispanic, and 10.5% missing racial and ethnic data), after excluding a subset of participants who were missing > 25% of items for one or more questionnaires (n = 19), or who displayed patterns of inconsistent responding on the PPI (who scored ≥3 SD above the mean on the PPI VRIN scale; n = 13).

An undergraduate sample was utilized in the current study, as the primary aim was to characterize the continuous score relations of subscales of differing inventories of psychopathy and normal range personality with facet scales of the TriPM, rather than to compare classifications of individuals as psychopathic or non-psychopathic based on cut-scores for differing inventories. Indeed, many of the scales administered in the current study were specifically developed for use in non-clinical settings (e.g., PPI, YPI). Further, although undergraduate students tend to display less severe levels of psychopathic traits as a whole than incarcerated individuals, the configurations of scores on psychopathy measures are largely consistent across community and forensic samples (Williams, Paulhus, & Hare, 2007). Further, self-report measures of psychopathy and normal-range personality share a great deal of psychometric overlap, as both provide information across a similar range of the latent traits assessed (Walton, Roberts, Krueger, Blonigen, & Hicks, 2008), suggesting considerable continuity between normal and abnormal measures and traits. Similarly, Sellbom and Phillips (2013) demonstrated that the patterns of external correlates of the TriPM were consistent across samples of undergraduates and female prisoners. Moreover, undergraduate samples are informative in and of themselves, as research using student samples may ultimately lead to a better understanding of how psychopathic traits are displayed in non-criminal settings and the conditions under which psychopathic traits are expressed in less antisocial forms (Lilienfeld, 1998).

Procedure

Prior to questionnaire administration, participants were informed of the nature of the study and provided written consent. Questionnaire data were collected in two waves, with the first 197 participants completing the protocol in-person via pencil-and-paper in groups of 5 to 20, and the remainder (n = 453) completing the questionnaires online using a secure internet-based survey system. Participants who completed the questionnaires in person versus electronically did not significantly differ in age, race, or gender. There were also no significant differences in TriPM total or scale scores for participants who completed the measures in person versus electronically (ps = .37 – .97). As such, data from the two waves of administration were combined for analyses. As compensation, participants received either $15, course credit, or a combination of the two.

Measures

Triarchic Psychopathy Measure (TriPM)

The TriPM (Patrick, 2010) is a non-proprietary 58-item inventory that yields scores on subscales of Boldness, Meanness, and Disinhibition, corresponding to constructs of the Triarchic Model, with subscales summed to yield a Total Psychopathy score.

The items of the TriPM Disinhibition and Meanness scales are from the Externalizing Spectrum Inventory (ESI; Krueger et al., 2007), a comprehensive index of traits and problem behaviors associated with disinhibitory (externalizing) psychopathology. The ESI’s 23 subscales load together on a general disinhibitory factor, with residual variances for some subscales defining separate subfactors reflecting callous-aggressive tendencies and substance abuse proneness. The TriPM Disinhibition scale comprises 20 items that strongly predict scores on the general disinhibitory factor of the ESI (r > .9; Patrick, Kramer, Krueger, & Markon, in press). The TriPM Meanness scale comprises 19 items selected to be strongly predictive (r > .8; Patrick et al., in press) of scores on the ESI callous-aggression factor, after controlling for moderate overlap (r = .45 in the present sample) with scores on the TriPM Disinhibition scale. The relatively high correlation between scores on the TriPM Meanness and Disinhibition scales is expected, given that both scales were derived from the same measurement model.

The third TriPM scale, Boldness, consists of 19 items from a new inventory (Patrick, Vaidyanathan, Benning, Hicks, & Kramer, 2012b) developed to refine conceptualization and measurement of the fearless-dominance dimension of the PPI. Items of this scale reflect tendencies toward fearlessness in three distinct domains: social efficacy (persuasiveness, social assurance, dominance), emotional stability (resiliency, confidence, optimism), and venturesomeness (courage, intrepidness, tolerance for uncertainty). Within the current sample, scores on TriPM Boldness correlated very strongly with scores on the fearless-dominance factor of the PPI (r = .82; see below), while showing only a modest positive association with TriPM Meanness (r = .23) and a weak negative association with TriPM Disinhibition scale (r = −.10). The modest correlation between TriPM Boldness and Meanness was somewhat lower than expected, given that fearless temperament is a hypothesized etiological substrate of both constructs (Frick & Marsee, 2006); however it is likely that other factors also contribute to dispositional meanness (Patrick et al., 2012a). Additional information regarding the psychometric properties, means, and standard deviations of the TriPM and other instruments administered in this study is provided in the [on-line] Appendices to this article.

Normal-Range Personality Measures

NEO-PI-R Antagonism

The NEO PI–R (Costa & McCrae, 1992) was developed to operationalize the Five-Factor Model (FFM) of personality. The 48 items of the NEO-PI-R Agreeableness scale were administered to current study participants, and reverse-scored to indicate higher Antagonism. A total Antagonism score was derived, along with scores on six lower-order facets: (lack of) Trust; (lack of) Straightforwardness; (lack of) Altruism; (lack of) Compliance; (lack of) Modesty; and (lack of) Tendermindedness.

35-item Multidimensional Personality Questionnaire (MPQ-35)

The MPQ (Tellegen, 1982) assesses personality in terms of 11 primary traits. The current study used a 35-item version developed for use in an epidemiological study of health and well-being across the life span (MIDUS-II; Javaras et al., 2012; www.midus.wisc.edu). The 35 items include representation of 10 of 11 trait scales of the full-length MPQ (Wellbeing, Social Potency, Achievement, Social Closeness, Stress Reaction, Aggression, Alienation, Control, Harm avoidance, Traditionalism; the 11^th scale, Absorption, is omitted).

Adult Psychopathy Inventories

Psychopathic Personality Inventory (PPI)

The PPI (Lilienfeld & Andrews, 1996; Lilienfeld & Widows, 2005), a 187-item inventory designed to measure the core personality features of psychopathy described in various historical accounts of the disorder, yields scores on eight empirically-derived subscales along with a composite total score. Factor analytic research also supports aggregation of scores on the Social Potency, Stress Immunity, and Fearlessness subscales of the PPI into a Fearless Dominance (FD) factor, and scores on the Machiavellian Egocentricity, Impulsive Nonconformity, Alienation [labeled Blame Externalization in the revised PPI-R; Lilienfeld & Widows, 2005], and Carefree Nonplanfulness subscales into an Impulsive-Antisociality (IA) factor. The remaining PPI subscale, Coldheartedness, does not load appreciably on either of these factors and thus is typically analyzed as a ‘stand-alone’ scale. The original PPI was used in the present study, because there is a more extensive research literature supporting the factor structure of the PPI compared to the PPI-R, as the revised scale was developed more recently.

Self-Report Psychopathy Scale-III (SRP-III)

The SRP-III (60 items; Paulhus, Hemphill, & Hare, 2009) was developed to index facets of the PCL-R via self-report. It yields a total psychopathy score and scores on four subscales: Callous Affect (e.g., lack of empathic concern for persons or animals), Interpersonal Manipulation (e.g., instrumental use of flattery), Erratic Lifestyle (e.g., impulsive decision making), and Criminal Tendencies (e.g., assault against a law enforcement officer).

Levenson’s Self-Report Psychopathy Scale (LSRP)

The 26-item LSRP (Levenson et al., 1995) was designed to assess “primary” and “secondary” subdimensions of psychopathy (Karpman, 1941). It yields a total psychopathy score along with scores on Primary and Secondary subscales, intended to reflect interpersonal-affective and impulsive-antisocial deviance features of psychopathy, respectively.

Juvenile Psychopathy Scales

Youth Psychopathic Traits Inventory (YPI)

The YPI (50 items; Andershed et al., 2002) was designed to assess the interpersonal, affective, and behavioral features of psychopathy in younger and older adolescents. It yields an overall psychopathy score and scores on 10 subscales, combinable into three broad factor scales: Grandiose-Manipulative, Callous-Unemotional, and Impulsive-Irresponsible.

Child Psychopathy Scale (CPS)

The CPS (Lynam, 1997) is a 50-item self-report version of a parent-rating scale for assessing psychopathic features in juveniles. It provides a total psychopathy score along with scores on 13 subscales, each composed of 2 to 7 items, intended to mirror the content of PCL-R items considered applicable to juveniles (i.e., omitting adult- or offender-oriented items such as promiscuity, multiple marriages, and parole/probation revocation).

Antisocial Process Screening Device (APSD)

The APSD (20 items; Frick & Hare, 2001) is a self-report version of a parent- and teacher-rating measure for assessing features of psychopathy in children, also modeled after the PCL-R. It yields a total psychopathy score and scores on Callous/Unemotionality and Impulsivity/Conduct Problems (I/CP) factors, with the latter divisible into Narcissism and Impulsivity subfactors. The CU Factor indexes shallow and restricted affect, lack of remorse or guilt, and lack of empathy. The I/CP Narcissism and I/CP Impulsivity subfactors index egocentric manipulativeness and proneness to boredom, rashness, and risk-taking, respectively.

Inventory of Callous-Unemotional Traits (ICU)

The 24-item Inventory of Callous-Unemotional Traits (ICU; Frick, 2004) was developed to refine conceptualization and measurement of “callous-unemotional traits” as indexed by the CU Factor of the APSD. The ICU yields a total score and three facet scores: Unemotional (e.g., restricted display of feelings or emotions), Callous (e.g., lack of guilt about wrongdoing), and Uncaring (e.g., insensitive to feelings of others). ²

Results

Correlational analyses were used to examine relations of TriPM total and subscale scores with adult and youth psychopathy inventories, and psychopathy-relevant personality trait measures, with presentation of results organized around a priori hypotheses. Relations with NEO-PI-R Antagonism and MPQ trait scale scores are presented in Table 1. Relations with adult and youth psychopathy inventories are presented in Tables 2 and 3, respectively. Along with simple bivariate (zero-order) correlations, each Table also includes standardized beta weights from regression analyses reflecting the unique contributions of the three TriPM scales to prediction of each criterion measure.³ Unless otherwise noted, results reported in the text reflect standardized beta weights from regression models, as these values allow for clearer delineation of distinctive relations of subscales of the TriPM with criterion measures (i.e., after controlling for overlap among the TriPM scales). To focus the presentation on findings most likely to be replicable, only associations of at least modest magnitude (r or β ≥ .20) are discussed in the text.

Table 1.

Relations between Normal-Range Personality Measures and Scores on the TriPM: Pearson Correlations and Regression Coefficients

	TriPM Total r	Boldness r (β)	Meanness r (β)	Disinhibition r (β)	Multiple R
Multidimensional Personality Questionnaire (35-item)
Social Potency	.34	.51a (.53)	.14 (−.04)	.04 (.12)	.52
Wellbeing	.09	.34 (.41)	−.11 (−.25)	−.06 (.10)	.41
Stress Reaction	.03	−.35a (−.34)	.10 (.06)	.33a (.26)	.46
Achievement	.04	.31 (.32)	−.07 (−.09)	−.19 (−.11)	.36
Harm Avoidance	−.34	−.31c (−.29)	−.24 (−.12)	−.14 (−.12)	.37
Social Closeness	.08	.13 (.22)	−.22a (−.31)	−.08 (.08)	.31
Aggression	.47	.01 (−.09)	.55a (.49)	.40 (.18)	.58
Alienation	.18	−.10 (−.13)	.22 (.17)	.27 c (.18)	.31
Control	−.43	−.19 (−.18)	−.33 (−.16)	−.35 c (−.30)	.44
Traditionalism	−.18	−.15 (−.14)	−.14 (−.08)	−.08 (−.06)	.19
NEO-PI-R Antagonism
Antagonism Total Score	.55	.19 (.11)	.55a (.44)	.37 (.19)	.58
(lack of) Trust	.24	−.12 (−.18)	.34c (.34)	.27 (.09)	.41
(lack of) Straightforwardness	.55	.31 (.27)	.47a (.30)	.34 (.24)	.55
(lack of) Altruism	.37	−.03 (−.09)	.44 b (.39)	.34 (.16)	.48
(lack of) Compliance	.49	.18 (.12)	.48a (.37)	.34 (.18)	.51
(lack of) Modesty	.30	.28 (.27)	.20c (.10)	.11 (.10)	.32
(lack of) Tendermindedness	.29	.16 (.09)	.32 a (.32)	.11 (−.03)	.34

Note: rs and βs > .13 are significant at p<.001; rs> .11 and βs > .07 are significant at p<.05; rs and βs ≥ .20 are highlighted in bold.

^aMagnitude of Pearson correlation significantly greater for hypothesized versus non-hypothesized scale (Steiger’s z ≥ 3.29, p<.001).

^bMagnitude of Pearson correlation significantly greater for hypothesized versus non-hypothesized scale (Steiger’s z ≥ 1.96, p<.05).

^cMagnitude of Pearson correlation for hypothesized scale does not significantly differ from non-hypothesized scale (Steiger’s z < 1.96, p>.05).

Table 2.

Relations between Adult Psychopathy Measures and Scores on the TriPM: Pearson Correlations and Regression Coefficients

	TriPM Total r	Boldness r (β)	Meanness r (β)	Disinhibition r (β)	Multiple R
Psychopathic Personality Inventory
PPI Total Score	.78	.52 (.51)	.57 (.26)	.49 (.43)	.79
PPI-FD	.57	.82a (.81)	.30 (.08)	.01 (.07)	.83
Social Potency	.45	.68a (.71)	.16 (−.07)	.05 (.16)	.70
Fearlessness	.58	.55a (.52)	.40 (.19)	.23 (.20)	.64
Stress Immunity	.30	.69a (.65)	.13 (.07)	−.24a (−.20)	.71
PPI-IA	.62	.06 (.04)	.55 (.30)	.67a (.55)	.73
Machiavellian Egocentricity	.62	.18 (.13)	.58b (.39)	.51 (.35)	.65
Impulsive Nonconformity	.53	.25 (.25)	.40 (.17)	.43c (.38)	.55
Alienation	.30	−.15 (−.14)	.29 (.14)	.49a (.41)	.51
Carefree Nonplanfulness	.34	−.11 (−.11)	.32 (.15)	.51a (.44)	.53
Coldheartedness	.33	.25 (.12)	.41a (.47)	.00 (−.20)	.47
Self-Report Psychopathy Scale-III
SRP-III Total Score	.69	.35 (.30)	.59 (.36)	.48 (.34)	.69
Interpersonal Manipulation	.56	.28c (.23)	.50 (.35)	.36 (.22)	.56
Callous Affect	.55	.32 (.22)	.55a (.47)	.26 (.06)	.59
Erratic Life Style	.66	.35 (.36)	.47 (.17)	.53c (.49)	.68
Criminal Tendencies	.44	.16 (.14)	.36 (.20)	.39c (.31)	.47
Levenson Self-Report Psychopathy Scale
LSRP Total	.57	.00 (−.04)	.57 (.38)	.61 (.44)	.69
Primary	.56	.18 (.10)	.56a (.42)	.42 (.24)	.60
Secondary	.39	−.18 (−.16)	.39 (.21)	.60a (.49)	.64

Note: rs and βs > .13 are significant at p<.001; rs> .11 and βs > .07 are significant at p<.05; rs and βs ≥ .20 are highlighted in bold.

^aMagnitude of Pearson correlation significantly greater for hypothesized versus non-hypothesized scale (Steiger’s z ≥ 3.29, p<.001).

^bMagnitude of Pearson correlation significantly greater for hypothesized versus non-hypothesized scale (Steiger’s z ≥ 1.96, p<.05).

^cMagnitude of Pea

rson correlation for hypothesized scale does not significantly differ from non-hypothesized scale (Steiger’s z < 1.96, p>.05).

Table 3.

Relations between Child Psychopathy Measures and Scores on the TriPM: Pearson Correlations and Regression Coefficients

	TriPM Total r	Boldness r (β)	Meanness r (β)	Disinhibition r (β)	Multiple R
Youth Psychopathic Traits Inventory
YPI Total	.61	.31 (.28)	.49 (.27)	.44 (.35)	.61
Grandiose Manipulation Factor	.51	.28c (.26)	.41 (.23)	.35 (.28)	.51
Dishonest Charm	.44	.22 (.21)	.35 (.19)	.32 (.26)	.44
Grandiosity	.34	.36 (.35)	.23 (.12)	.09 (.07)	.40
Lying	.44	.07 (.06)	.39 (.22)	.45 (.36)	.50
Manipulation	.49	.27 (.25)	.40 (.23)	.33 (.25)	.50
Callous/Unemotional Factor	.51	.33 (.24)	.50a (.44)	.20 (.04)	.55
Callousness	.38	.19 (.08)	.44 (.46)	.15 (−.07)	.46
Remorselessness	.45	.19 (.13)	.44 (.34)	.29 (.15)	.47
Unemotionality	.48	.44 (.37)	.39 (.30)	.12 (.03)	.53
Impulsive/Irresponsible Factor	.55	.15 (.21)	.36 (.03)	.62a (.63)	.66
Thrill-Seeking	.45	.27 (.30)	.28 (.05)	.37 (.38)	.48
Impulsiveness	.44	.07 (.14)	.28 (−.01)	.56 (.58)	.58
Irresponsibility	.43	.03 (.07)	.30 (.05)	.57 (.56)	.58
Inventory of Callous-Unemotional Traits
ICU Total Score	.39	.10 (−.01)	.48a (.48)	.22 (.00)	.48
Unemotional	.12	.03 (−.08)	.25a (.36)	−.04 (−.21)	.31
Callous	.35	.09 (.02)	.37b (.31)	.26 (.12)	.39
Uncaring	.35	.03 (−.05)	.41b (.37)	.28 (.11)	.42
Antisocial Process Screening Device
APSD Total Score	.58	.11 (.09)	.52 (.32)	.55 (.42)	.64
Narcissism	.39	.10c (.06)	.37 (.26)	.33 (.22)	.42
Callous-Unemotionality	.38	.02 (−.04)	.41b (.35)	.33 (.17)	.45
Impulsivity	.42	.05 (.08)	.32 (.10)	.49a (.46)	.51
Child Psychopathy Scale
CPS Total Score	.38	.06 (.06)	.33 (.16)	.40 (.34)	.44
Glibness	.30	.20 (.33)	.15 (−.03)	.17 (.23)	.36
Lack of Guilt	.27	.26 (.21)	.24 (.19)	.05 (−.01)	.32
Poverty of Affect	.12	−.18 (−.20)	.17 (.13)	.27 (.19)	.33
Manipulation	.31	.14 (.16)	.21 (.05)	.28 (.28)	.33
Callousness	.28	.03 (.00)	.28 (.21)	.25 (.16)	.32
Untruthfulness	.32	.11 (.09)	.27 (.15)	.27 (.22)	.33
Parasitic Lifestyle	.31	.07 (.09)	.23 (.07)	.33 (.31)	.36
Behavioral Dyscontrol	.13	−.26 (−.27)	.20 (.15)	.34 (.25)	.43
Unreliability	.19	−.09 (−.06)	.16 (.04)	.33 (.30)	.33
Failure to Accept Responsibility	.34	.15 (.17)	.24 (.07)	.32 (.30)	.37
Lack of Planning	.24	.02 (.02)	.21 (.12)	.26 (.21)	.28
Impulsiveness	.23	−.05 (.00)	.15 (−.02)	.38 (.39)	.38
Boredom Susceptibility	.41	.18 (.20)	.28 (.08)	.37 (.36)	.44

Note: rs and βs > .13 are significant at p<.001; rs> .11 and βs > .07 are significant at p<.05; rs and βs ≥ .20 are highlighted in bold.

^aMagnitude of Pearson correlation significantly greater for hypothesized versus non-hypothesized scale (Steiger’s z ≥ 3.29, p<.001).

^bMagnitude of Pearson correlation significantly greater for hypothesized versus non-hypothesized scale (Steiger’s z ≥ 1.96, p<.05).

^cMagnitude of Pearson correlation for hypothesized scale does not significantly differ from non-hypothesized scale (Steiger’s z < 1.96, p>.05).

Hypothesis 1

Consistent with prediction, overall NEO-PI-R Antagonism scores were related most strongly to TriPM Meanness (β = .44), with Disinhibition and Boldness showing more modest independent associations. TriPM Meanness also exhibited the strongest unique association with each facet of Antagonism (βs = .30–.44), aside from lack of Modesty, which was predicted most strongly by Boldness (β = .27). The reversed Straightforwardness facet was related uniquely to all three TriPM facets, indicating that each contains content relevant to deceptiveness. Additionally, TriPM Disinhibition evidenced modest Pearson correlations with the reversed Trust, Altruism, and Compliance facets of Antagonism, and there was no significant difference between the magnitude of the correlation between TriPM Meanness and Disinhibition with the Trust facet.

Hypothesis 2

Overall TriPM scores were associated positively with MPQ Aggression and Social Potency, and negatively with Control and Harm Avoidance (Verona et al., 2001). TriPM Boldness was associated with elevations on all facets of MPQ Positive Emotionality (i.e., Social Potency, Wellbeing, Achievement, Social Closeness; Tellegen, 2011), and low levels of Stress Reaction and Harm Avoidance. Meanness was related primarily to low Social Closeness and Wellbeing, and high Aggression. Disinhibition was associated with low Control and high Stress Reaction (as well as with high Alienation and Aggression at the zero-order level). In general, predicted correlations were significantly greater in magnitude than non-predicted correlations; however, the magnitude of the correlations for TriPM Boldness and Meanness with Harm Avoidance did not significantly differ from one another, nor did the correlations for Meanness and Disinhibition with Control or Alienation.

Hypothesis 3

Consistent with hypothesis, overall TriPM scores correlated strongly with total scores on other omnibus psychopathy inventories (rs = .38 to .78, ps<.001). Correlations with the PPI and SRP-III were high enough (rs = .78 and .69) to conclude that the TriPM indexes largely the same construct within the domain of self-report (Campbell & Fiske, 1959). In contrast, correlations between TriPM and LSRP total scores were somewhat more modest (.57). Correlations with the child-oriented psychopathy measures (ASPD, CPS) were smaller (.58 and .38), with the adolescent-oriented YPI measure falling in between (r = .61).

As a further approach to evaluating the extent to which omnibus psychopathy measures assess psychopathic traits in a similar way across individual participants, additional continuous correspondence analyses were performed (Gass & Gonzalez, 2003). Total scores on the TriPM and adult and child psychopathy inventories were transformed to T-scores (M = 50, SD =10). Following transformation, T-score differences were computed between total scores on the TriPM and total scores on the various psychopathy measures. Average T-score differences for the 6 scales ranges from −.15 (CPS) to .06 (APSD), with standard deviations ranging from 6.56 (PPI) to 11.01 (CPS). Next, we calculated for each omnibus psychopathy measure, the percentage of individuals whose total score matched their total score on the TriPM within +/− 10 T-score points. Across the six adult and youth psychopathy measures, the percent of cases in which total scores on these measures fell within +/− 10 T-score points of their total score on the TriPM ranged from 72.1% for the LSRP to 86.6% for the PPI; M = 79.2%, SD = 5.4). Thus, while the majority of participants were classified as similarly high on psychopathic traits using the TriPM and other omnibus psychopathy self-report inventories (i.e., within 1 standard deviation) there is some mismatch in absolute rank-order status, which may be attributable to the degree to which each of the Triarchic phenotypes is indexed in various instruments, which is explored in the remaining hypotheses.

Hypothesis 4

Consistent with the conception of psychopathy as a callous-aggressive variant of externalizing psychopathology, TriPM Meanness and Disinhibition each contributed uniquely to prediction of total scores on each of the adult psychopathy measures (PPI, SRP-III, LSRP; βs ≥ .34, ps < .001). The magnitude of unique prediction was largely equivalent for the SRP-III and LSRP (see Table 2), indicating that Meanness and Disinhibition are represented to a similar degree in differing adult psychopathy inventories. Disinhibition contributed somewhat more to unique variance in PPI total scores than Meanness. TriPM Meanness and Disinhibition also contributed to prediction of total scores on the YPI and APSD (βs ≥ .27, ps < .001), but in each case here Disinhibition contributed more strongly than Meanness. Likewise, total scores on the CPS were most strongly related to TriPM Disinhibition (β=.34). The implication is that self-report child psychopathy scales contain greater representation of content related to general externalizing proneness than callous-unemotionality.

At the level of specific content scales, TriPM Meanness was most strongly associated with subscales of other psychopathy inventories indexing constructs such as callousness, manipulation, remorselessness, and unemotionality. In contrast, TriPM Disinhibition was most strongly associated with subscales indexing constructs such as impulsivity, irresponsibility/unreliability, boredom proneness/thrill-seeking, lack of planfulness, and criminal deviancy.

Hypothesis 5

At the zero-order level, ICU total scores were significantly related to both TriPM Meanness and Disinhibition; however, in accordance with the primary hypothesis, Meanness emerged as the sole unique predictor of ICU scores (β = .48) when all three TriPM facets were entered concurrently as predictors – indicating that the zero-order association between Disinhibition and ICU was attributable to the overlap of Meanness and Disinhibition. Meanness also emerged as the strongest unique predictor of the ICU Callous, Unemotional, and Uncaring facets (βs = .31–.37). Notably, after accounting for the overlap between Meanness and Disinhibition, scores on TriPM Disinhibition were inversely related to scores on the ICU Unemotional facet, presumably reflecting the elevated negative affectivity component of externalizing proneness (Krueger et al., 1996; Sher & Trull, 1994). TriPM Boldness showed negligible relations with ICU total or facet scores.

Hypothesis 6

Psychopathy inventories varied in the extent to which they indexed the adaptive features of psychopathy encompassed by Boldness. Consistent with hypothesis, Boldness contributed most strongly (over and above Disinhibition and Meanness) to prediction of total scores on the PPI (β = .51), with the next strongest predictive association evident for SRP-III total scores (β = .30; see Table 2). Notably, Boldness did not contribute significantly to prediction of LSRP total scores. Among the youth-oriented psychopathy inventories, YPI was the only measure for which TriPM Boldness contributed uniquely to prediction of total scores (β= .28; see Table 3). The correlation between Boldness scores and total scores on the PPI was significantly greater in magnitude than the correlations between Boldness and total scores on any of the other psychopathy self-report measures (Steiger’s z = 3.70 – 10.75, p<.001), which is to be expected, as subscales associated with the Fearless-Dominance Factor of the PPI served as one referent for the TriPM Boldness scale.

Hypothesis 7

We predicted that observed relations of TriPM Boldness with inventories modeled after the PCL-R (SRP-III, YPI, APSD, CPS) would be traceable mainly to items reflecting the interpersonal features of psychopathy. Partial support was obtained for this hypothesis, with youth measures conforming more to expectation than adult measures. Consistent with prediction, the unique contribution of TriPM Boldness to scores on the CPS was strongest for the Glibness subscale (β = .33), although some unique contribution of Boldness was also evident for subscales reflecting Lack of Guilt, Boredom Susceptibility, Behavioral Dyscontrol (−), and Poverty of Affect (−). Unique contributions of TriPM Boldness were evident as expected for three ‘interpersonal’ subscales of the YPI (Grandiosity, Manipulation, Dishonest Charm; βs = .21–.35); however, contributions were also evident for two of the YPI ‘affective’ and ‘impulsive-antisocial’ subscales (Unemotionality, Thrill-Seeking).

Contrary to prediction, none of the APSD subscales, including Narcissism, evidenced associations with TriPM Boldness, either in simple correlations or regression analyses. While, as noted above, TriPM Boldness related uniquely to overall scores on the SRP-III, its unique predictive contribution was not attributable solely or primarily to intersection with the Interpersonal Manipulation facet. Rather, Boldness contributed similarly to prediction of SRP-III Callous Affect facet (βs = .22) as to Interpersonal Manipulation (β = .23), and even more so to prediction of SRP-III Erratic Lifestyle (βs =.36). Additionally, SRP-III Manipulation showed equivalent or stronger relations with TriPM Meanness and Disinhibition (βs = .22 and .35, respectively) than with TriPM Boldness. This lack of clear mapping of TriPM scales onto SRP-III subscales likely reflects in part the moderate (.4–.6) interrelations among the SRP-III scales; this scale overlap contributes to the excellent internal consistency of the SRP-III (α = .93), but limits the separation of distinguishable facets of psychopathy within this measure.

Hypothesis 8

Consistent with the fact that the PPI-FD subscales served as referents for the model of fear and fearlessness upon which the TriPM Boldness scale was developed, and demonstrated in prior findings (Sellbom & Phillips, 2013; Marion et al., 2013; Stanley et al., 2013), PPI-FD scores showed significant zero-order associations with TriPM Boldness and Meanness; however, when Boldness, Meanness, and Disinhibition were entered concurrently as predictors of PPI-FD, Boldness remained the only significant predictor (β = .81), indicating that the zero-order association between Meanness and PPI-FD was attributable to overlap of Meanness with Boldness. TriPM Boldness also displayed its strongest associations with scores on the three subscales of PPI-FD (Social Potency, Stress Immunity, Fearlessness). The magnitude of relationship between TriPM Boldness and PPI-FD indicates that these measures index largely equivalent constructs.

Scores on PPI-IA were significantly related to both TriPM Meanness and Disinhibition, with Disinhibition accounting more for prediction (β = .55 versus β = .30). TriPM Boldness was unrelated to PPI-IA, either at the zero-order level or when entered with the other TriPM scales in a regression model. TriPM Disinhibition evidenced markedly stronger unique associations than TriPM Meanness with 3 of the 4 subscales comprising PPI-IA, Carefree Nonplanfulness, Alienation (alternatively labeled Blame Externalization), and Impulsive Nonconformity. The fourth PPI-IA scale, Machiavellian Egocentricity, was related equivalently to TriPM Meanness as to Disinhibition. These findings indicate that the observed zero-order association of TriPM Meanness with PPI-IA is attributable largely to its association with PPI Machiavellian Egocentricity.

Coldheartedness, the only subscale of the PPI not associated with either of its Factors, was related at the zero-order level to both TriPM Meanness and Boldness; however, when the three TriPM scales were entered together as predictors of Coldheartedness, Meanness emerged as the strongest significant predictor (β = .47), with TriPM Disinhibition contributing modestly in the negative direction, indicating that the zero-order association for Boldness was attributable to its overlap with Meanness. Thus, the two subscales of the PPI that relate most to scores on TriPM Meanness are Coldheartedness and Machiavellian Egocentricity.

As shown in Table 2, TriPM Boldness, Meanness, and Disinhibition each exhibited significant associations with PPI Fearlessness at the zero-order level. When examined jointly as predictors, TriPM Boldness showed the strongest unique relationship with PPI Fearlessness (β = .52), with associations lower for Disinhibition and Meanness (βs = .20 and .19). By contrast, TriPM Boldness and Disinhibition contributed in opposing directions (+ versus −) to prediction of scores on PPI Stress Immunity. Notably, Stress Immunity correlated only modestly with TriPM total scores (r = .30), which appears consistent with the literature more broadly, in which trait anxiety is often associated weakly or negligibly with psychopathy total scores. This has led some researchers to question the role of trait anxiety (and its operationalization vis-à-vis the Stress Immunity subscale of the PPI) in the construct network of psychopathy (Visser, Ashton, & Pozzebon, 2012). However, in the current sample, the Multiple R for the three TriPM scales predicting PPI Stress Immunity was dramatically higher (.71) than the zero-order association between Stress Immunity and TriPM total scores, highlighting the importance of operationalizing facets of psychopathy separately in order to optimize relations with criterion measures (Hicks & Patrick, 2006).

Hypothesis 9

LSRP total and subscale scores were largely unrelated to TriPM Boldness. In contrast, TriPM Meanness and Disinhibition showed significantly associations with scores on the LSRP as a whole and both of its subscales. Consistent with prediction, TriPM Meanness demonstrated a larger unique relationship with scores on the Primary scale of the LSRP whereas Disinhibition showed a larger unique relationship with scores on the Secondary scale; however, Meanness and Disinhibition remained significant predictors of each LSRP subscale when entered together in regression analyses. Thus, the LSRP Primary and Secondary scales each reflect a confluence of Meanness and Disinhibition, rather than indexing one or the other selectively.

Discussion

The present results demonstrate that alternative self-report psychopathy inventories index the constructs of the Triarchic Model as operationalized by the subscales of the TriPM to varying degrees. Clarification of commonalities and distinctions among measures in this domain should be of help in reconciling the sometimes conflicting findings in the literature that arise from the use of varied instruments for assessing psychopathy.

Some limitations of the current study warrant mention. One is that participants consisted exclusively of undergraduate students. While some evidence exists for the validity of the TriPM scales in prisoners (Sellbom & Phillips, 2013; Stanley et al., 2013; see also Venables & Patrick, 2012), future research ought to examine the construct validity of the TriPM in additional populations, including clinical and forensic samples of other types and using demographically diverse participants from the general community, which include greater diversity in terms of age, race, IQ, and levels of adaptive functioning. Along these lines, more advanced structural analyses of the TriPM scales are warranted to further evaluate their measurement properties in relation to other measures developed to index the constructs of the Triarchic model (Slaney & Maraun, 2008; Slaney, Storey, & Barnes, 2011). Additionally, the current study suffers from mono-method bias, as it relies exclusively on self-report measures of psychopathic traits. Due to shared method variance, the relations of the TriPM with criterion measures in the current study are necessarily higher than would be expected for criteria in other domains, such as interview-based diagnosis, observed behavior, or physiological reactivity. Future research would benefit from inclusion of psychopathy-relevant criterion variables from non-self-report domains.

Notwithstanding these limitations, the current results provide new insights into contents represented in alternative existing measures of psychopathy and have implications for research and applied clinical assessment, particularly with regard to non-forensic samples.

Disinhibition

The Disinhibition scale of the TriPM was represented strongly in each of the adult and youth psychopathy inventories. Supporting its content validity, scores on TriPM Disinhibition were most strongly associated with indices of the behavioral deviance aspect of psychopathy, as indicated by elevations on subscales tapping impulsivity, irresponsibility, carelessness, thrill-seeking, and antisocial behavior. With respect to adult psychopathy inventories, Disinhibition and Meanness contributed similarly to unique variance in total psychopathy scores. This coincides with the notion that disinhibitory features alone are not sufficient to diagnose psychopathy; interpersonal and affective features (vis-à-vis meanness and/or boldness) must also be present (Patrick, Venables, & Drislane, 2013).

In contrast, TriPM Disinhibition contributed more strongly than Meanness to prediction of total scores on child and adolescent psychopathy inventories. This could indicate that assessment instruments operationalize psychopathy differently across the lifespan, with disinhibitory features (i.e., impulsivity, irresponsibility) emphasized more strongly in childhood conceptions. Notably, aside from the Youth Psychopathic Traits Inventory (YPI), TriPM Boldness was not strongly represented in the child-oriented inventories. This finding can be interpreted in differing ways. One possibility is that boldness may be less essential to defining psychopathy in childhood, with callous-unemotional tendencies (a.k.a. meanness) representing the more prototypic expression of core affective-interpersonal traits early in life. Alternatively, it may be that boldness is expressed in children with psychopathic traits, but lacks effective representation in existing youth psychopathy inventories, perhaps because the child psychopathy literature has focused predominantly on samples exhibiting conduct problems. If so, future research may benefit from ascertaining whether characteristics like social efficacy, stress resiliency, and venturesomeness (reflecting boldness) may contribute to the emergence or expression of psychopathy in children or adolescents as in adults.

Meanness

As with Disinhibition, TriPM Meanness was strongly represented in all adult and child psychopathy self-report inventories. Supporting its content validity, scores on Meanness were most strongly associated with scales indexing coldheartedness/callous affect, unemotionality, Machiavellianism, manipulation, and remorselessness. Further, consistent with the assertion that antagonism is central to defining psychopathy (Widiger & Lynam, 1998), Meanness contributed uniquely to prediction of scores on the majority of subscales of the SRP-III, LSRP, YPI, and APSD, as well as to total scores on these inventories. Additionally, among the TriPM scales, Meanness emerged as the strongest predictor of specific personality trait variables considered relevant to psychopathy, including NEO-PI-R Antagonism and callous-unemotional traits as measured by the ICU.

Boldness

The role of boldness (Fearless Dominance) in the conceptualization of psychopathy has been at the forefront of recent debates in the literature (Lilienfeld et al., 2012; Miller & Lynam, 2012; Patrick et al., 2013). Some researchers contend that boldness should not be considered a core element of psychopathy given its association with indices of adaptive functioning (Miller & Lynam, 2012). Other researchers have argued that adaptive correlates are in fact consistent with Cleckley’s (1976) original description of psychopathy as a masked disorder in which severe behavioral deviance goes hand in hand with an outward appearance of robust mental health (Lilienfeld et al., 2012). Further, the emotional stability associated with phenotypic boldness appears to be one of the features that most strongly distinguishes psychopathy from Antisocial Personality Disorder and other externalizing conditions, which are typically associated with heightened negative affectivity and comorbid internalizing psychopathology (Patrick et al., 2013; Krueger, 1999).

In the current study, TriPM Boldness was associated with indices of maladjustment (manipulativeness, callous affect, erratic lifestyle, dishonesty, grandiosity/lack of modesty, guiltlessness) as well as with measures of adaptive function (charm, stress immunity, social potency, wellbeing, achievement). For example, TriPM Boldness was the strongest predictor of the lack of Modesty facet of NEO-PI-R Antagonism. This finding is consistent with the notion that Boldness captures, at least in part, the interpersonal features of psychopathy indexed by facet 1 of the PCL-R, which includes the items glibness/superficial charm and grandiosity. In view of this, Boldness as operationalized in the TriPM cannot be conceived of as purely adaptive. This may particularly be the case when boldness is accompanied by high levels of meanness or disinhibition. Research evaluating interactive effects of differing psychopathy facets on criterion measures of interest is needed in order to clarify how observable clinical manifestations are moderated by alternative configurations of these distinctive facets.

Implications for Assessment of Psychopathy

The present study demonstrates the utility of the Triarchic Psychopathy Measure as a rubric for comparing various psychopathy assessment instruments in non-forensic samples. Although each instrument purports to index a common psychological phenomenon (albeit, each reflecting a particular model or set of theoretical models regarding the psychopathy construct), our findings indicate that alternative self-report inventories entail somewhat differing operationalizations of psychopathy. Some instruments, such as the Child Psychopathy Scale (CPS), assess psychopathy with a stronger emphasis on general externalizing proneness (i.e., prominent representation of disinhibition). Other inventories, like Levenson’s Self-Report Psychopathy Scale (LSRP), represent a blending of general externalizing proneness with callous-aggression (i.e. strong representation of both disinhibition and meanness). Finally, instruments like the Psychopathic Personality Inventory (PPI) and the Self-Report Psychopathy Scale-III (SRP-III) incorporate aspects of adaptive functioning in addition to externalizing problems in their assessment of psychopathy (i.e. prominent representation of boldness along with meanness and disinhibition).

An understanding of these differences across assessment instruments is likely to be beneficial in evaluating findings in the literature, as the external correlates of separate facets of psychopathy are known to be quite divergent. For example, in the current study, differing facets as indexed by the TriPM evidenced differential associations with measures of stress immunity and emotional resilience: Whereas Meanness was largely unrelated to Stress Immunity (β = .07), Boldness was associated with markedly elevated Stress Immunity (β = .65), and Disinhibition was associated with diminished Stress Immunity (β = −.20). In light of such divergences, the relative representation of each of these facets in a particular assessment inventory will necessarily impact the relations between that inventory and criterion measures of interest -- including clinically-important outcomes such as treatment responsiveness, violent recidivism, and suicide. This highlights the importance of assessing psychopathy as a multifaceted construct, in order to capture unique predictive relations for its distinct symptomatic components.