Clarifying associations between psychopathy facets and personality disorders among offenders

(1)

Tilburg University

Clarifying associations between psychopathy facets and personality disorders among

offenders

Klipfel, Kristen M.; Garofalo, C.; Kosson, D.S.

Published in:

Journal of Criminal Justice

DOI:

10.1016/j.jcrimjus.2017.09.002

Publication date:

2017

Document Version

Peer reviewed version

Link to publication in Tilburg University Research Portal

Citation for published version (APA):

Klipfel, K. M., Garofalo, C., & Kosson, D. S. (2017). Clarifying associations between psychopathy facets and personality disorders among offenders. Journal of Criminal Justice, 53, 83-91.

https://doi.org/10.1016/j.jcrimjus.2017.09.002

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal

Take down policy

(2)

Clarifying Associations between Psychopathy Facets and

Personality Disorders among Offenders

Kristen M. Klipfel1, Carlo Garofalo2, and David S. Kosson1

1_{Department of Psychology, Rosalind Franklin University of Medicine and Science, North Chicago}

(IL), United States 2_{Department of Developmental Psychology, Tilburg University, Tilburg, the}

Netherlands

Abstract

Purpose—This study examined bivariate, unique, and multivariate associations between

psychopathy facets and other Personality Disorders (PDs).

Method—76 incarcerated males were assessed with clinical interviews assessing psychopathy

and DSM-5 PDs. Canonical Correlation Analysis (CCA) was used to examine multivariate associations between dimensional scores of psychopathy facets and other PDs.

Results—Preliminary analyses of bivariate and partial associations revealed that much of the

covariation between psychopathy and PD traits reflected shared variance among psychopathy facets and among PD traits. After controlling for the shared variance, unique relationships were limited to positive relationships between Narcissistic PD and interpersonal facet and between Paranoid PD and antisocial facet ratings. Canonical Correlation Analysis results yielded two pairs of functions that explained the shared variance between psychopathy and PDs. In the first pair of functions, elevations on the interpersonal and antisocial facets were associated with symptoms of Paranoid, Narcissistic, Histrionic, and Antisocial PDs. In the second pair of functions, high levels of the antisocial facet and low levels of the interpersonal facet were related to Borderline PD.

Conclusion—Results suggest that associations between psychopathy and DSM-5 PDs go beyond

established links with Antisocial and Narcissistic PDs to include associations with Histrionic, Borderline, and Paranoid PDs.

Keywords

psychopathy; personality disorders; canonical correlation analysis; DSM-5; shared and unique variance

Correspondence concerning this article should be addressed to Kristen M. Klipfel, Department of Psychology, Rosalind Franklin University of Medicine and Science, 3333 Green Bay Road, North Chicago, IL 60064. kmklipfel@gmail.com.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our

customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

These data, with the exception of the results pertaining to Canonical Correlation Analysis (CCA), were orally presented at the sixth biennial meeting of the Society for the Scientific Study of Psychopathy. The initial findings, as well as additional analyses (i.e., CCA),

HHS Public Access

Author manuscript

J Crim Justice. Author manuscript.

Author Manuscript

(3)

Psychopathy is a personality disorder (PD) characterized by a constellation of affective, interpersonal, and behavioral features, including but not limited to: lack of empathy, guilt, or remorse; shallow affectivity; interpersonal manipulation; impulsivity and irresponsibility; and persistent antisocial tendencies (Cleckley, 1976; Hare & Neumann, 2005). Psychopathy is most often clinically assessed using the Psychopathy Checklist-Revised (PCL–R; Hare, 2003). Whereas the original studies examining the factor structure of the Psychopathy Checklist (PCL) and the PCL–R yielded a two-factor conceptualization of psychopathy (Hare, et al., 1990), more recent advances in psychopathy research have highlighted the value of further subdividing Factors 1 and 2 into four facets or lower order dimensions (Hare & Neumann, 2008). Factor 1, commonly described as the core affective and interpersonal personality components of psychopathy features, has been subdivided into more narrow-band interpersonal and affective facets. Factor 2, commonly described as the antisocial lifestyle (or social deviance) component of psychopathy, has been subdivided into more narrow-band facets reflecting an impulsive and irresponsible lifestyle, as well as early, persistent, and versatile antisocial tendencies.

Although there is evidence that psychopathy can be conceptualized at multiple levels (the overall disorder, the higher-order factors, or the more specific facets), a focus on the lower-order dimensions of psychopathy has several advantages. First, the use of lower lower-order dimensions allows researchers to study components of psychopathy with greater precision. Not only do the facet level models provide better explanations for the relationships between scores on the PCL-R items, they also provide information about more homogeneous constructs than the higher-order factors (Hare & Neumann, 2008; Smith, McCarthy, & Zapolski, 2009). In addition, research using the facets has demonstrated distinct patterns of relationships for the two facets comprising each higher-order factor (Graham, Kimonis, Wasserman, & Kline, 2012; Hoppenbrouwers, Neumann, Lewis, & Johansson, 2015; Vitacco, Neumann, & Jackson, 2005; Walsh, Swogger, & Kosson, 2009; Walters, Knight, Grann, & Dahle, 2008; for a review, see Hare & Neumann, 2008).

Psychopathy is one of the most empirically well-validated disorders of personality pathology (Miller, Gaughan, & Pryor, 2008), and it has significant forensic and clinical implications for understanding some forms and functions of antisocial behavior, as well as for designing and implementing interventions (DeLisi, 2009). Psychopathy is not yet listed as a distinct PD in the main body of the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, APA, 2013). Not only is psychopathy not currently identified as a distinct PD in DSM-5 (although it is proposed as a specifier for Antisocial PD in a supplementary section), but the empirical literature on psychopathy and other PDs has grown rather separately, with the former of chief interest in criminal justice and forensic systems and the latter of primary interest in clinical settings (Fossati, Pincus, Borroni, Munteanu, & Maffei, 2014).

Even though the diagnostic criteria for Antisocial PD were originally developed to capture the clinical construct of psychopathy (Robins, Tipp, Przybeck, 1991), there is substantial evidence that PCL-R-assessed psychopathy and DSM-5 Antisocial PD represent two distinct (but related) operationalizations of the syndrome, with some different characteristics and associated consequences, as well as with different prevalence, prognoses, and underlying

Author Manuscript

(4)

mechanisms (Gregory et al., 2012; Kosson, Lorenz, & Newman, 2006; Riser & Kosson, 2013). These differences primarily reflect the fact that criteria for Antisocial PD emphasize the lifestyle and antisocial components of psychopathy, whereas the construct of

psychopathy also includes a complex pattern of specific interpersonal features and emotional dysfunctions, which are not required to meet a diagnosis of Antisocial PD (Ogloff, 2006). Accordingly, studies that have adopted the two-factor model of psychopathy have

consistently reported lower correlations between ratings of Antisocial PD and PCL-R Factor 1 psychopathy than between ratings of Antisocial PD and PCL-R Factor 2 psychopathy (Crego & Widiger, 2014; Harpur, Hare, & Hakstian, 1989).

The relationship between psychopathy and Narcissistic PD has also received substantial attention. Kernberg (1992) was one of the first to suggest that narcissism, malignant narcissism, Antisocial PD, and psychopathy exist on the same continuum, demonstrating overlapping features among these constructs. In this account, narcissism was distinguished from psychopathy by a less pronounced antisocial lifestyle and by a relatively more intact ability to engage in significant relationships with others (Gunderson & Ronningstam, 2001; Kernberg, 1992). This perspective is also consistent with Cleckley’s (1976) seminal

description of psychopathy, which included pathological egocentricity (that is, a prototypical narcissistic trait) as a defining feature of psychopathy. As a result, items capturing

grandiosity, entitlement, and arrogance have been incorporated in many measures used to examine psychopathic traits (e.g., the PCL-R [Hare, 2003], and the Psychopathic Personality Inventory [Lilienfeld & Andrews, 1996]). More recently, researchers have argued that there are relationships between psychopathy and narcissism (e.g., Hart & Hare, 1998; Miller et al., 2010), and evidence suggests Narcissistic PD is associated with ratings on both of the factors of the PCL–R, though the correlation is typically larger for Factor 1 than for Factor 2 (Fossati et al., 2005).

Although psychopathic features have most often been linked to Antisocial and Narcissistic PDs, fewer studies address relationships between psychopathy and other PDs. This lack of research seems unfortunate because understanding the similarities and differences between psychopathy and other PDs is important for improving the current diagnostic system (e.g., reducing redundancy across disorders) and improving treatment programs, by identifying what is unique to psychopathy and what is shared with other PDs (Fossati et al., 2005). In some studies that have examined these relationships, preliminary findings have linked psychopathy to Paranoid PD (Blackburn, 2007; Blackburn & Coid, 1998; Warren & Burnette, 2013) and other Cluster A PDs, which include Schizotypal (Ragsdale & Bedwell, 2013; Rogers, Jordan, & Harrison, 2007; Warren & Burnett, 2013) and Schizoid (Warren & Burnett, 2013) PDs, as well as to Borderline (Miller et al., 2010) and Histrionic PDs (Coid et al., 2009). Table 1 includes all prior studies that have examined zero-order correlations between dimensional scores on PCL/PCL–R psychopathy and dimensional scores on other clinically measured PDs. The most replicated associations have been between ratings of Antisocial PD and ratings on all four psychopathy facets, as well as between ratings of Narcissistic PD and ratings on the interpersonal and affective facets of psychopathy. Some consistency has also emerged linking Histrionic PD with the interpersonal and affective components (and, to a lesser extent, with the lifestyle and antisocial components) of psychopathy, as well as linking symptoms of Borderline and Paranoid PD with the lifestyle

Author Manuscript

(5)

and antisocial components (and, to a lesser extent, with the affective and interpersonal components) of psychopathy. Consistent with the evidence for substantial comorbidity among the PDs, the evidence of links between psychopathy components and eight of the ten PDs included in the DSM-5 suggests the possibility that substantial shared variance among PDs may contribute to these relationships. Similarly, the moderate to large correlations among psychopathy facet scores suggest the possibility that shared variance among the psychopathy dimensions also contributes to these relationships.

To date, only one published study has reported analyses designed to remove the variance shared by scores on PDs and ratings on psychopathy facets. Coid et al. (2009) used multiple regression analyses to examine unique associations between ratings on each psychopathy facet and symptoms of other PDs after controlling for the variance associated with other PDs, as well as for scores on the other psychopathy facets, substance abuse, psychosis, age, and sex. They reported unique associations between ratings of Antisocial PD and ratings on all facets of psychopathy, between ratings of Narcissistic PD and of the interpersonal and affective facets, between ratings of Histrionic PD and of the interpersonal and lifestyle facets, and between ratings of Schizoid PD and of the affective facet.

Findings like these are important for elucidating which relationships between psychopathy and other PDs are unique, and these results suggest that several of the relationships between specific psychopathy components and specific PD symptoms are robust. However, because Coid et al. (2009) did not report the zero-order correlations between psychopathy facet ratings and other PD symptoms, it is not possible to evaluate whether these unique relationships were also present at the zero-order level or whether they reflect novel associations that emerged only after removing the variance shared among psychopathy facets or PDs. More concretely, when analyses control for scores that correlate substantially with predictor variable scores, the resulting regression coefficients refer to residualized scores on both PDs and psychopathy facets, which do not correspond to the way these syndromes appear in nature (e.g., see Lynam, Hoyle, & Newman, 2006). Moreover, because Coid et al. controlled for scores on other variables that share substantial variance with psychopathy factors (e.g., substance use, sex of participant, etc.), they removed additional substantive variance from scores on psychopathy facets and other PDs. Finally, only studies that report both zero-order and unique associations can indicate which relationships between psychopathy and other PDs reflect specific associations and which reflect the shared variance among the different PDs and among the psychopathy facets.

One way to integrate results regarding bivariate associations and unique contributions between two sets of variables is to adopt a multivariate approach that takes into account the shared variance within each set at the same time, without removing what is shared from the calculation of the single coefficients. One such approach is Canonical Correlation Analysis (CCA). The multivariate approach of CCA has methodological and conceptual advantages when examining the relationships between psychopathy and other PDs (Courville & Thompson, 2001; Thompson, 1991; Sherry & Henson, 2005); however, to our knowledge, it has not been used to examine these associations. First, consistent with evidence regarding the substantial comorbidities among PDs, CCA examines PDs on a continuum rather than as separate categories. Second, it allows for simultaneous comparisons between multiple

Author Manuscript

(6)

predictors and multiple dependent variables without inflating Type 1 error. More

specifically, this approach considers the covariation within each variable set to estimate the psychopathy facet ratings and the other PD symptom ratings that contribute to the

correlation between the two composite sets of variables. Further, by considering the overlap between psychopathy facets, as well as the overlap among other PDs, it seems plausible that constellations of psychopathic traits (e.g., a combination of interpersonal and affective traits) could be related to a constellation of PD symptoms (e.g., a combination of narcissistic and histrionic features). However, zero-order correlations and multiple regression analyses only allow for the examination of associations between single variables; they do not allow for the examination of associations between constellations or sets of variables. In short, CCA allows researchers to investigate relationships between multivariate composites (i.e., linear combinations) of variables. Moving beyond the focus on psychopathy facets and PDs in isolation, this approach enables us to understand the covariation between constellations of psychopathic traits and constellations of PD traits. Considering the high overlap between psychopathy facets and between PDs, this approach seems to appropriately resemble clinical reality.

The current study was designed to provide a more comprehensive examination of the relationships between the components of psychopathy and symptoms of other PDs. We conducted comprehensive assessments of both psychopathy and DSM-5 PDs, using valid clinical measures of both psychopathy and DSM-5 PDs in a sample of incarcerated males. Personality disorders were assessed dimensionally in keeping with current trends in personality disorder (e.g., Livesley & Jang, 2000; Wright et al., 2012) and clinical forensic research (e.g., DeLisi & Vaughn, 2015; Ullrich, Borkenau, & Marneros, 2001). We examined the zero-order relationships between psychopathy facet ratings and other PD symptom scores, as well as the unique relationships between psychopathy facet ratings and other PD symptom scores after controlling for scores on both the other psychopathy facets and the other PDs. Finally, we sought to extend prior knowledge by examining which variables contributed meaningfully to the latent associations between psychopathy facets and PDs using CCA. This combination of analytic methods allowed us to systematically examine the differential associations between psychopathy facets and other PDs.

Method

Participants

The study sample was drawn from a larger sample of 298 inmates assessed on the PCL–R for which there were ratings provided by both an interviewer and an observer. However, the sample of this study consisted of 76 participants that were administered clinical measures of both psychopathy and PDs (see below). All participants were males detained at a county jail in the Midwest of the United States. Participants ranged in age from 18–43 years (M = 26.4, SD = 6.7). The sample consisted primarily of African Americans (63.6 percent), followed by 23.4 percent European Americans, 10.4 percent Latino Americans, and 2.6 percent

identified as Other. The number of formal schooling years ranged from six to sixteen (M = 11.4, SD = 1.6), and 48.1 percent of participants had either completed a high school education or earned a General Equivalency Diploma (GED).

Author Manuscript

(7)

Measures

Psychopathy Checklist-Revised (PCL–R)—The PCL–R (Hare, 2003) is a 20–item

instrument used to assess traits and behaviors associated with psychopathy. Ratings on the PCL–R have been shown to have excellent validity and reliability, as well as psychometric procedures (Hare, et al., 1990). Each item is rated on a 3-point scale (absent, inconsistent/ subthreshold, present) based on information obtained from a semi-structured interview, behavioral observations, and collateral file material. Total PCL–R scores provide valid measures of psychopathy in European American (Hare, 1996), African American (Cooke, Kosson, & Michie, 2001), and Latino American male offenders (Sullivan, Abramowitz, Lopez, & Kosson, 2006). Factor analytic studies across diverse samples (Neumann, Hare, & Pardini, 2014) demonstrate that the PCL–R items may be grouped into four correlated components, namely the affective, interpersonal, lifestyle, and antisocial facets. In the present study, trained graduate students conducted the semi-structured interviews before gathering file information to complete psychopathy ratings. In the current sample, dual ratings were available for 6 participants (7.9%), as they were the only individuals that had trained observers present for both the PCL–R and the International Personality Disorders Examination (IPDE). The inter-rater reliability for these ratings (computed using intraclass correlation coefficient, ICC) was .94. For the full sample of 298 inmates, the interrater agreement was .95, indicating excellent interrater agreement. Internal consistency alpha’s (α) and mean inter-item correlations (MIC) for the PCL-R facets were: interpersonal, α = . 71, MIC = .39; affective, α = .71, MIC = .38; lifestyle, α = .60, MIC = .23; antisocial, α = . 61, MIC = .24.

The International Personality Disorders Examination (IPDE)—The IPDE

(Loranger, 1988) is a semistructured clinical interview that assesses the personality disorders listed in the International Statistical Classification of Diseases and Health Related Problems, 10th Revision (ICD-10; WHO, 1993) and in the fourth edition of the Diagnostic and Statistical Manual of Mental Disorders, (DSM-IV; American Psychiatric Association, 1994) classification systems. Items reflecting personality disorder criteria are grouped into six thematic headings: work, self, interpersonal relationships, affects, reality testing, and impulse control. The IPDE covers all of the criteria for the 10 Axis II Personality Disorders in the text revision of the DSM-IV, namely: Paranoid, Schizoid, Schizotypal, Histrionic, Borderline, Antisocial, Narcissistic, Dependent, Avoidant, and Obsessive–Compulsive. Of note, all of these criteria remained completely unchanged in the DSM-5 (American Psychiatric Association, 2013). Interrater reliability of IPDE diagnoses is generally good (median kappa = .73), as is test-retest reliability (median = .87; Blanchard & Brown, 1998). In the current study, dual ratings for the IPDE interviews were available for 22 participants (29.0%), with ICCs ranging from .834 (for Paranoid PD) to .951 (for Obsessive–Compulsive PD). Internal consistency α’s and MIC for the IPDE scales were: Paranoid PD, α = .76, MIC = .31; Schizoid PD, α = .42, MIC = .12; Schizotypal PD, α = .71, MIC = .23; Histrionic PD, α = .54, MIC = .13; Antisocial PD, α = .47, MIC = .13; Narcissistic PD, α = .79, MIC = .29; Dependent PD, α = .48, MIC = .13; Avoidant PD, α = .63, MIC = .20;

Obsessive-Compulsive PD, α = .67, MIC = .21.1

Author Manuscript

(8)

Procedure

Individuals were chosen at random from a sample of eligible male inmates at a county jail in the Midwest. They were contacted via telephone and given information about the study. Male inmates were eligible if they were: 1) 18 years of age or older; 2) not taking psychotropic medications; 3) booked and awaiting a trial for a felony or misdemeanor or already convicted and serving a sentence; and 4) spoke English and could read at a 4th grade level. Participants completed an interview designed to allow scoring of the PCL–R, followed by a review of available jail records, and then they completed the IPDE on a separate day. Doctoral level graduate students formally trained on the PCL–R and IPDE conducted all interviews. Two different interviewers conducted the PCL–R and IPDE interviews, such that IPDE raters were blind to ratings on the PCL–R. The University’s Institutional Review Board (IRB) approved all of the study’s methods and materials. All participants provided written consent before proceeding with the study.

Data Analytic Strategy

Descriptive statistics were computed for all study variables. Bivariate and unique associations were investigated using Pearson product-moment correlations and multiple regression analyses, respectively. We then conducted a CCA using the four psychopathy facets as predictors and the 10 PD variables as criterion variables to evaluate the

multivariate shared relationship between the two sets of variables (i.e., psychopathy facets and PDs). As discussed above, the purpose of the CCA was to determine which variables contributed most to the associations between psychopathy facet ratings and PD dimensional scores, as reflected by their loadings on the synthetic (i.e., latent) variables that explained a significant proportion of the variance shared by the two sets. Canonical correlation analysis is a multivariate method that allows researchers to identify the variables that account for the most substantial proportion of shared variance between the two sets (Sherry & Henson, 2005; Thompson, 1991). This analysis produces pairs of synthetic (i.e., latent) functions or variates that maximize the correlation between the two sets of variables and provides coefficients that indicate which variables in each set contribute most to each of these synthetic functions – without removing the variance shared within each set. For each model estimated, Wilks’s λ represents the unexplained variance of the overall model, and 1 – λ provides the effect size of the overall model. Specifically, each structure coefficient (rs)

represents a measure of the effect size for the relation between each variable and the synthetic function generated by the two composite sets of variables. Notably, the rs’s

produced in a CCA are resistant to collinearity and suppression effects and are akin to the structure coefficients used in factor analysis. Therefore, in interpreting each rs, we adopted

the conventional rule of thumb of considering effect sizes equal to or greater than .45 as meaningful. For increased reliability, canonical correlation requires a ratio of about ten cases for every independent variable (Tabachnick & Fidell, 2007). In the present study, the independent variables were the four psychopathy facets, suggesting a sample size of at least 40 participants was needed to ensure adequate power to perform the CCA.

1Internal consistency estimates could not be computed for the Borderline PD scale of the IPDE because it consists of items that are weighted differently to produce a scale score, and no indications on how to estimate its internal consistency are provided in the IPDE manual (Loranger, 1988).

Author Manuscript

(9)

Results

Zero-Order Correlations

Table 2 shows descriptive statistics and bivariate associations for all study variables. When examining the zero-order correlations among PDs and PCL–R total and facet scores, dimensional scores for most PDs (all except Schizoid, Avoidant, and Obsessive–Compulsive PDs) correlated significantly with ratings on traits associated with psychopathy. Scores on Borderline, Histrionic, Antisocial, Narcissistic, Schizotypal, and Paranoid PDs correlated significantly with PCL–R total scores. Scores on Histrionic and Narcissistic PDs were correlated with ratings on the interpersonal facet of psychopathy. In contrast, no PDs were significantly correlated with ratings on the affective facet, all r’s < .19, ps > .10. Scores on Borderline, Histrionic, Schizotypal, and Dependent PDs were correlated with ratings on the lifestyle facet, all r’s > .25, p’s < .03. Scores on Borderline, Histrionic, Antisocial,

Narcissistic, Paranoid, and Schizotypal PDs were correlated with ratings on the antisocial facet, all r’s > .28, p’s < .02.

Unique Associations

To examine unique associations between scores on PD symptoms and psychopathy facet ratings while simultaneously controlling for the shared variance in each domain, we used residualized scores for each psychopathy facet as dependent variables in multiple regression analyses, with all PD dimensional scores entered as independent variables. Results indicated that residualized scores on the interpersonal facet were positively associated with

Narcissistic PD (β = .58, p < .001) and negatively associated with Borderline PD (β = −.56, p < .001). The overall model explained approximately 41% of the variance in interpersonal facet ratings, R2 = .41, F(10, 62) = 4.26, p < .001. Residualized scores on the antisocial facet were also uniquely related to Paranoid PD scores (β = .42, p < .05); this model significantly explained roughly 34% of the variance in antisocial facet ratings, R2 = .34, F(10, 62) = 3.16, p < .01. Finally, the models predicting residualized scores on the affective and lifestyle facets of psychopathy were not significant, F(10, 62) = 1.86, and F(10, 62) = 1.38, respectively, all ps > .05.

Canonical Correlation Analysis

The CCA produced four pairs of canonical variates (i.e., functions), which is, as expected, the same as the number of variables in the smaller of the two variable sets (i.e., psychopathy facets). In particular, the four pairs of functions had squared canonical correlations of . 44, .38, .17, and .12, for each successive function. Overall, the full model across all functions was statistically significant, Wilk’s λ = .25, F(40, 229.37) = 2.50, p < .001, and indicates a substantial portion (i.e., roughly 75%) of the variance shared between the two variable sets was explained.

Dimension reduction analysis showed that the second variate was also significant, as the cumulative effect of Functions 2 through 4 was statistically significant, Wilk’s λ = .45, F(27, 178.79) = 2.07, p < .001. However, the cumulative effect of Functions 3 and 4, as well as the effect of Function 4 alone, were not significant, Wilk’s λ = .73, F(16, 124) = 1.33, and Wilk’s λ = .88, F(7, 63) = 1.23, respectively (all ps > .05). Because the third and fourth

Author Manuscript

(10)

variates did not explain a statistically significant amount of shared variance between the two sets of variables, only the first two functions were interpreted.

The first pair of variates was linked by a canonical correlation of .66, demonstrating a large effect size. Specifically, the two sets of variables shared approximately 44 percent of the variance . After extracting the first function, the second variate explained roughly 38 percent of the remaining variance and also revealed a large effect size (Rc = .

61). The second pair of variates explained approximately 25 percent of the total variance in the model.

Table 3 shows standardized canonical function coefficients, structure coefficients, and squared structure coefficients for the first two pairs of functions, as well as the

communalities across the functions for each variable. Inspection of structure coefficients (rs)

revealed the psychopathy facets that showed the largest loadings on the first variate were the interpersonal and antisocial facets. All coefficients were negative. Among the PD variables, the largest loadings on the first function were for Paranoid, Histrionic, Narcissistic, and Antisocial PDs. The negative coefficients that denote the rs of all these variables indicate

that symptoms for these PD variables were positively related to levels of the interpersonal and antisocial facets. Of note, the effect sizes were large for Paranoid, Histrionic, and Narcissistic PD variables and were only moderate for the Antisocial PD variable. Regarding the second pair of functions, the main predictors were again the antisocial and interpersonal facets (though with positive and negative signs, respectively). Of note, the effect size was relatively larger for the antisocial facet. Among the PDs, only Borderline PD made a substantial positive contribution to this function. As such, this function appears to highlight a positive association between Borderline PD and the antisocial facet of

psychopathy and a negative association between Borderline PD and the interpersonal facet of psychopathy. Notably, the patterns of results based on the squared structure coefficients for the two pairs of functions were consistent with the corresponding standardized canonical function coefficients (akin to Beta weights in multiple regression analysis), as well as with the communality indices. A graphical depiction of the canonical correlation results (limited to the first two pairs of functions) is presented in Figure 1.

Discussion

Notwithstanding the well-established overlap among PDs, there has been little empirical research about the extent to which specific PDs are associated with psychopathic traits. It is important to examine these associations to better understand the extent to which the construct of psychopathy can be understood as related to combinations of existing PD symptoms and the extent to which psychopathy has unique features, as well as to be able to evaluate the relevance of evidence-based interventions designed for other PDs for clinical work with offenders with psychopathic traits. In an attempt to advance current knowledge on this topic, we analyzed and compared bivariate and multivariate associations between dimensional scores of both psychopathy and DSM-5 PDs, assessed via widely well-validated clinical measures.

Author Manuscript

(11)

At the zero-order level, almost all of the PD dimensional scores correlated significantly with traits associated with psychopathy; only scores on Schizoid, Avoidant, and Obsessive– Compulsive PDs did not correlate with either the PCL–R total or facet ratings. Scores on all four Cluster B PDs, as well as Paranoid PD and Schizotypal PD, exhibited a substantial number of correlations with PCL–R total and facet ratings. The correlations between psychopathy ratings and scores on Paranoid PD and Schizotypal PD are also consistent with a few previous studies examining relationships between psychopathy and Cluster A PDs (e.g., Blackburn, 2007; Ragsdale & Bedwell, 2013; Raine, 1992; Warren & Burnette, 2013). The associations between ratings on psychopathy dimensions and scores on the four Cluster B PDs are consistent with prior findings demonstrating high comorbidity between

psychopathy and Cluster B PDs (Blackburn, 2007; Miller et al., 2010; Rogers, Jordan, & Harrison, 2007). Such correlations are also consistent with the perspective that psychopathy has more in common with Cluster B PDs than with PDs in the other two clusters

(Huchzermeier et al., 2007). This finding is also consistent with a basic trait perspective, which posits that much of the comorbidity between different PDs can be understood in light of common underlying traits. Specifically, it has been argued that much of the overlap between psychopathic traits and Cluster B PDs may be partly explained in terms of underlying traits related to antagonism and disinhibition (Lynam & Widiger, 2001).

The differences between the correlations for facet ratings versus overall psychopathy ratings suggest several of the relationships are specific to some facets of psychopathy. Notably, scores on only two of the PDs, namely Histrionic and Narcissistic PDs, were associated with the interpersonal features of psychopathy, and there were no significant zero-order

correlations with the affective facet. Conversely, there were several correlations with the lifestyle and antisocial components of psychopathy. Scores on Borderline, Histrionic, Schizotypal, and Dependent PDs were significantly correlated with ratings on the lifestyle facet. Scores on all four Cluster B PDs, as well as Paranoid and Schizotypal PDs, were significantly correlated with ratings on the antisocial facet. In general, the associations of these two facets with Cluster B PDs, and the associations with Paranoid, Schizotypal, and Dependent PD symptoms, are consistent with previous findings and seem to suggest that a greater overlap with PDs exist for the behavioral features of psychopathy. As a whole, these results seem consistent with those of the studies reviewed in Table 1, in that the associations between psychopathy facets and PD symptoms in prior studies were primarily between scores on Cluster B PDs and ratings on the lifestyle and antisocial facets of psychopathy. However, a comparison of zero-order and unique relationships showed that very few of these relationships reflected unique relationships between psychopathy facets and specific PDs. First, when examining unique associations between residualized scores on the

interpersonal facet and PDs, only Narcissistic PD showed a significant positive contribution, which may reflect commonality in antagonism. Interestingly, there was also a unique negative association with Borderline PD features that emerged. Further, none of the PD symptom scores were uniquely related to lifestyle facet ratings, and only scores on Paranoid PD remained significantly associated with ratings on the antisocial facet. The unique relationship with Paranoid PD suggests that the antisocial facet of psychopathy does not simply capture behavioral indices of antisociality as operationalized in the DSM-5 criteria for Antisocial PD. This finding appears consistent with theory and research suggesting that

Author Manuscript

(12)

the antisocial facet of psychopathy, at least as assessed with the PCL–R, may reflect more than just disinhibition and antisocial behavior. This finding is also consistent with studies linking psychopathic traits with hostile attribution bias, which is also a defining feature of Paranoid PD (Vitale, Newman, Serin, & Bolt, 2005). Another interesting finding is that Histrionic and Antisocial PD were not uniquely related to any psychopathy facets, despite the strong zero-order correlations between these PD symptoms and psychopathy ratings, indicating that relationships between the DSM-5 conceptualizations of Histrionic and Antisocial PD and psychopathy primarily reflect the variance shared among the psychopathy facets.

To our knowledge, this was the first study to adopt a multivariate approach to the association between psychopathy and other PDs, in order to clarify the different findings obtained from zero-order and unique associations. Using CCA, we found evidence for a substantial association between psychopathy facets and symptoms of other PDs, as the canonical correlation that linked the variates representing each set of variables was large in magnitude. Specifically, our model explained 75 percent of the shared variance between psychopathy and other PDs. Our results also revealed two pairs of functions that

meaningfully captured the shared variance between psychopathy and PDs. High levels of both the interpersonal and antisocial facets contributed to the first psychopathy function. Based on these loadings, this function appears to fit with conceptualizations of psychopathic personality that emphasize manipulativeness and grandiosity, as well as aggression and antisociality (i.e., elevations on both core psychopathic features and behavioral features). The interpersonal and antisocial facets also defined the second psychopathy function, though with opposite signs, such that high levels of the antisocial facet and low levels of the interpersonal facet characterized the remaining shared variance with other PDs. Therefore, this second function suggests an important distinction between the combination of interpersonal and antisocial features of psychopathy (Function 1) and elevation of the antisocial (i.e., behavioral) features in the absence of the core interpersonal features of psychopathy.

Taken together, these results suggest that an exclusive focus on specific components of psychopathy can lead to an inaccurate understanding of their associations with other PDs. Whereas the variates produced by the composite set of psychopathy facets indicated that some of the associations between psychopathy and other PDs are likely due to a

constellation of psychopathic traits comprising the interpersonal and antisocial components, regression and correlations indicate associations reflect specific relationships with the unique variance in one facet of psychopathy and associations reflect the shared variance among the different components of psychopathy.

An in-depth examination of the structure coefficients revealed a more specific picture regarding the PD symptoms associated with these constellations of psychopathic traits. The first function revealed a substantial contribution of Paranoid, Histrionic, Antisocial, and Narcissistic PD symptoms to the variance shared with both the interpersonal and antisocial features of psychopathy. Contrary to expectations based on zero-order correlations, yet in line with results from multiple regression analyses, Antisocial PD made a relatively weak contribution to the model, indicating that the shared variance associated with both

Author Manuscript

(13)

psychopathy and DSM-5 PDs may be more strongly related to other PDs – notably, Paranoid, Histrionic, and Narcissistic PDs – than to Antisocial PD. Taken together, these findings appear to suggest that the overlap between psychopathy and other PDs may be partly due to traits related to antagonism (i.e., paranoid, histrionic, and narcissistic traits), and less so to traits related to disinhibition (i.e., Antisocial PD). Of note, these results suggest that by focusing on a constellation of PDs, rather than focusing on a single PD, we can obtain a better understanding of the continuities between psychopathy and other PDs. For instance, it is possible that some of the interpersonal and antisocial features of

psychopathy may be partly understood as encompassing processes shared with symptoms of Histrionic (e.g., impression management), Antisocial (e.g., lying), Narcissistic (e.g., grandiosity), and Paranoid (e.g., hostility) PDs – which appear to share an antagonistic component. In addition, this pattern of associations suggests that ratings on the antisocial facet of psychopathy are strongly associated with PDs that are not commonly related to crime, including Histrionic PD. It is worth noting that no unique associations between Histrionic and Antisocial PDs and the interpersonal or antisocial facets of psychopathy emerged in multiple regression analyses. Taken together, this suggests that controlling for shared variance among PD symptoms and among PCL-R facets left only relatively trivial and nonsignificant coefficients linking the residual variance in Histrionic and Antisocial PDs with the residual variance in the interpersonal and antisocial facets of psychopathy. Thus, these associations would likely have been overlooked had we not used a multivariate technique like CCA.

On the other hand, an examination of the second function revealed a positive relation between symptoms of Borderline PD and the antisocial facet of psychopathy, as well as a negative relation between symptoms of Borderline PD and the interpersonal facet. These results suggest that there is also a positive relationship between early, persistent, and versatile antisocial tendencies and the pattern of general behavioral dysregulation associated with Borderline PD, which may reflect a disposition toward dysregulated emotions,

behaviors, and interpersonal relationships. It should also be noted that the associations involving Borderline PD were not evident when the variance shared among PDs was statistically removed in multiple regression analyses, a phenomenon quite similar to what we observed for Histrionic and Antisocial PDs. Finally, it should be emphasized that the second pair of functions was obtained after removing the variance explained by the first pair of functions. As such, results concerning the second pair of functions should be interpreted with caution (Lynam et al., 2006).

Multivariate results were also consistent with associations at the zero-order level in that the affective facet of psychopathy seems to represent a component of psychopathy that is not closely related to symptoms of other PDs. These findings are consistent with findings from other studies indicating that the affective component of psychopathy is less related to symptoms of substance use disorders than the other components of psychopathy (e.g., Walsh, Allen, & Kosson, 2007) and with the perspective that the affective traits of psychopathy may be relatively more specific to psychopathy than some of the other components of psychopathy, at least as assessed by the PCL–R.

Author Manuscript

(14)

The findings for the lifestyle component of psychopathy are somewhat different. Although the CCA findings suggest that the lifestyle facet is not one of the primary sources of the shared variance between psychopathy and indices of other PDs, the dramatic discrepancy between the zero-order correlations and the partial correlations for the lifestyle facet

indicates that the apparent relationships between the lifestyle component of psychopathy and symptoms of other PDs primarily reflect variance shared between this component and the other components of psychopathy. Overall, these findings contribute to an increasing body of literature highlighting the importance of investigating facet-level associations between psychopathy and external correlates (Hoppenbrouwers et al., 2015; Vitacco et al., 2005). A comparison between bivariate and multivariate results also indicated that some of the other associations reported at the zero-order level (e.g., those involving the lifestyle facet of psychopathy and symptoms of Borderline, Schizotypal and Dependent PDs) were likely due to the shared variance within each set of variables. For instance, it is possible that ratings on Schizotypal PD are related to psychopathy in part because they share some features with Paranoid PD (Blackburn, 2007; Blackburn & Coid, 1998; Hare, 1985; Warren & Burnette, 2013). Likewise, zero-order correlations between ratings of Borderline PD and the

interpersonal facet may have been due to the tendency to manipulate others, which is part of both the interpersonal facet of psychopathy and of Borderline PD. However, manipulation is a characteristic that is also present in several other PDs (e.g., Histrionic, Narcissistic, and Antisocial PDs). When this commonality across PDs is taken into account, Borderline PD is negatively related to a residualized index of the interpersonal facet, which suggests it may only be indirectly related to the interpersonal facet because of its shared variance with other PDs, especially Cluster B PDs. Finally, the multivariate approach adopted in this study also revealed a number of meaningful relationships that were obscured when examining unique associations (i.e., holding constant the shared variance among psychopathy facets and among PDs). Most notably, scores on Histrionic PD and Paranoid PD correlated positively with ratings on both the interpersonal and antisocial facets. This pattern of findings is consistent with the recommendation that information regarding bivariate, unique, and multivariate associations should be integrated to obtain a comprehensive understanding of the links between interrelated sets of variables (Lynam et al., 2006).

This study has several limitations. First, it contained a relatively small sample size, though acceptable considering the special nature of the sample and the difficulties inherent in conducting two relatively long clinical assessments based on semi-structured interviews with each participant. However, the current sample size ensured adequate power for CCA (Tabachnick & Fidell, 2007). Second, the mean PCL–R total score was high, and therefore generalizations to samples with lower average levels of psychopathy should be made with caution. Additionally, the study sample was comprised only of male offenders, as the number of eligible women incarcerated at the institution was far less than the number of eligible men. As a result, the generalizability of these findings cannot be extended to female offenders, as PD symptoms may manifest differently across the two sexes (e.g., Weizmann-Henelius, Viemerö, & Eronen, 2004); thus, an important goal for future research is to examine the covariation between psychopathy and PDs among women. Finally, recent clinical and research advances have supported the existence of two different phenotypical expressions of Narcissistic PD (i.e., vulnerable and grandiose narcissism; e.g., Cain, Pincus,

Author Manuscript

(15)

& Ansell, 2008) and suggests that each phenotypical expression may relate differently to psychopathic traits (Miller et al., 2010). However, it has been posited that the DSM-based conceptualization might only capture the grandiose aspects of Narcissistic PD

(Ronningstam, 2010). Further investigations adopting a multidimensional assessment of Narcissistic PD are greatly needed.

With these limitations in mind, this study adds to the current literature by indicating that symptoms of Paranoid and Histrionic PDs may have more relevance to psychopathy than previously noted. Future studies examining the overlap among psychopathy and other PDs would do well to consider the similarities among Paranoid and Histrionic PDs and psychopathy. This study also provides convincing evidence that the interpersonal and antisocial facets make a joint contribution to the shared variance between psychopathy and other PDs. Similarly, associations between Borderline, Histrionic, and Narcissistic PDs and the antisocial facet of psychopathy and unique associations between Paranoid PD and antisocial facet ratings contribute to a growing body of evidence suggesting that the antisocial facet does not merely reflect Antisocial PD. Further, findings suggest that there is something unique about some components of psychopathy (i.e., affective and lifestyle traits), which is not captured by the current PD nosology in the DSM-5. Finally, results indicate that a multivariate approach can greatly improve our understanding of the association between psychopathy and PDs, highlighting the importance of focusing on a constellation of psychopathic traits and PDs, rather than on psychopathic traits and PDs in isolation.

Acknowledgments

This research was funded by the National Institute of Mental Health Grant, MH090169, awarded to David S. Kosson. We gratefully acknowledge the assistance of Rick Riddle, G. Mark McCorley, and the staff of the Lake County Jail for their support of this study, and Amy Bagley, Chelsea Brieman, Michael Brook, Allison Brown Erica Christian, Sarah Hampton, Jami Mach, Christine Meltzer, Katie Montry, Stephanie Smith, Elizabeth Sullivan, Marc Swogger, and Zach Walsh who assisted with data collection and data entry.

References

American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th. Washington, DC: Author; 1994.

American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 5th. Washington, DC: Author; 2013.

Blackburn R. Personality disorder and antisocial deviance: Comments on the debate on the structure of the psychopathy checklist-revised. Journal of Personality Disorders. 2007; 21(2):142–159.

[PubMed: 17492918]

Blackburn R, Coid JW. Psychopathy and the dimensions of personality disorder in violent offenders. Personality and Individual Differences. 1998; 25(1):129–145.

Blanchard JJ, Brown SB. Structured diagnostic interview schedules. Assessment. 1998; 4:97–130. Cain NM, Pincus AL, Ansell EB. Narcissism at the crossroads: Phenotypic description of pathological

narcissism across clinical theory, social/personality psychology, and psychiatric diagnosis. Clinical Psychology Review. 2008; 28(4):638–656. [PubMed: 18029072]

Cleckley, HM. The mask of sanity. 5th. St. Louis, MO: Mosby; 1976.

Coid J, Yang M, Ullrich S, Roberts A, Moran P, Bebbington P, Brugha T, Jenkins R, Farrell M, Lewis G, Singleton N, Hare R. Psychopathy among prisoners in England and Wales. International Journal of Law and Psychiatry. 2009; 32(3):134–141. [PubMed: 19345418]

Author Manuscript

(16)

Cooke DJ, Kosson DS, Michie C. Psychopathy and ethnicity: Structural, item, and test generalizability of the psychopathy checklist-revised (PCL–R) in Caucasian and African American participants. Psychological Assessment. 2001; 13(4):531. [PubMed: 11793896]

Courville T, Thompson B. Use of structure coefficients in published multiple regression articles: β is not enough. Educational and Psychological Measurement. 2001; 61:229–248.

Crego C, Widiger TA. Psychopathy and the DSM. Journal of Personality. 2014; 83:665–677. [PubMed: 25039353]

DeLisi M. Psychopathy is the Unified Theory of Crime. Youth Violence and Juvenile Justice. 2009; 7(3):256–273.

DeLisi M, Vaughn MG. Ingredients for criminality require genes, temperament, and psychopathic personality. Journal of Criminal Justice. 2015; 43(4):290–294.

Fossati A, Beauchaine TP, Grazioli F, Carretta I, Cortinovis F, Maffei C. A latent structure analysis of diagnostic and statistical manual of mental disorders, fourth edition, narcissistic personality disorder criteria. Comprehensive Psychiatry. 2005; 46(5):361–367. [PubMed: 16122536] Fossati A, Pincus AL, Borroni S, Munteanu AF, Maffei C. Are pathological narcissism and

psychopathy different constructs or different names for the same thing? A study based on Italian nonclinical adult participants. Journal of Personality Disorders. 2014; 28(3):394–418. [PubMed: 24511898]

Graham N, Kimonis ER, Wasserman AL, Kline SM. Associations among childhood abuse and psychopathy facets in male sexual offenders. Personality Disorders: Theory, Research, and Treatment. 2012; 3(1):66.

Gregory S, Simmons A, Kumari V, Howard M, Hodgins S, Blackwood N. The antisocial brain: Psychopathy matters: A structural MRI investigation of antisocial male violent offenders. Archives of General Psychiatry. 2012; 69(9):962–972. [PubMed: 22566562]

Gunderson J, Ronningstam E. Differentiating antisocial and narcissistic personality disorder. Journal of Personality Disorders. 2001; 15:103–109. [PubMed: 11345846]

Hare RD. Comparison of procedures for the assessment of psychopathy. Journal of Consulting and Clinical Psychology. 1985; 53(1):7–16. [PubMed: 3980831]

Hare RD. Psychopathy a clinical construct whose time has come. Criminal Justice and Behavior. 1996; 23(1):25–54.

Hare, RD. Manual for the revised psychopathy checklist. 2nd. Toronto, Ontario: Multi-Health Systems; 2003.

Hare RD, Harpur TJ, Hakstian AR, Forth AE, Hart SD, Newman JP. The revised Psychopathy Checklist: Reliability and factor structure. Psychological Assessment: A Journal of Consulting and Clinical Psychology. 1990; 2(3):338.

Hare RD, Neumann CS. Structural models of psychopathy. Current Psychiatry Reports. 2005; 7(1):57– 64. [PubMed: 15717988]

Hare RD, Neumann CS. Psychopathy as a clinical and empirical construct. Annu Rev Clin Psychol. 2008; 4:217–246. [PubMed: 18370617]

Harpur TJ, Hare RD, Hakstian AR. Two-factor conceptualization of psychopathy: Construct validity and assessment implications. Psychological Assessment: A Journal of Consulting and Clinical Psychology. 1989; 1(1):6.

Hart, S., Hare, RD. Association between psychopathy and narcissism: Theoretical views and empirical evidence. In: Ronningstam, EF., editor. Disorders of narcissism: Diagnostic, clinical, and

empirical implications. Washington, DC: American Psychiatric Press; 1998. p. 415-436.

Hildebrand M, de Ruiter C. PCL–R psychopathy and its relation to DSM-IV Axis I and II disorders in a sample of male forensic psychiatric patients in the Netherlands. International Journal of Law and Psychiatry. 2004; 27(3):233–248. [PubMed: 15177992]

Hoppenbrouwers SS, Neumann CS, Lewis J, Johansson P. A latent variable analysis of the Psychopathy Checklist–Revised and behavioral inhibition system/behavioral activation system factors in North American and Swedish offenders. Personality Disorders: Theory, Research, and Treatment. 2015; 6(3):251.

Huchzermeier C, Geiger F, Bruß E, Godt N, Köhler D, Hinrichs G, Aldenhoff JB. The relationship between DSM-IV cluster B personality disorders and psychopathy according to Hare’s criteria:

Author Manuscript

(17)

Clarification and resolution of previous contradictions. Behavioral Sciences & the Law. 2007; 25(6):901–911.

Kernberg, OF. Aggression in personality disorders and perversions. New Haven, CT: Yale University Press; 1992.

Kosson DS, Lorenz AR, Newman JP. Effects of comorbid psychopathy on criminal offending and emotion processing in male offenders with antisocial personality disorder. Journal of Abnormal Psychology. 2006; 115(4):798. [PubMed: 17100537]

Lilienfeld SO, Andrews BP. Development and preliminary validation of a self-report measure of psychopathic personality traits in noncriminal population. Journal of personality assessment. 1996; 66(3):488–524. [PubMed: 8667144]

Livesley WJ, Jang KL. Toward an empirically based classification of personality disorder. Journal of Personality Disorders. 2000; 14:137–151. [PubMed: 10897464]

Loranger, AW. Personality disorder examination (PDE) manual. Yonkers, New York: DV Communications; 1988.

Lynam DR, Hoyle RH, Newman JP. The perils of partialling cautionary tales from aggression and psychopathy. Assessment. 2006; 13(3):328–341. [PubMed: 16880283]

Lynam DR, Widiger TA. Using the five-factor model to represent the DSM-IV personality disorders: an expert consensus approach. Journal of abnormal psychology. 2001; 110(3):401. [PubMed: 11502083]

Miller JD, Dir A, Gentile B, Wilson L, Pryor LR, Campbell WK. Searching for a vulnerable Dark Triad: Comparing Factor 2 psychopathy, vulnerable narcissism, and borderline personality disorder. Journal of Personality. 2010; 78:1530–1564.

Miller JD, Gaughan ET, Pryor LR. The Levenson Self-Report Psychopathy Scale: An examination of the personality traits and disorders associated with the LSRP factors. Assessment. 2008; 15:450– 463. [PubMed: 18480502]

Neumann CS, Hare RD, Pardini DA. Antisociality and the construct of psychopathy: Data from across the globe. Journal of Personality. 2014 [PubMed: 25181550]

Ogloff JR. Psychopathy/antisocial personality disorder conundrum. Australian and New Zealand Journal of Psychiatry. 2006; 40(6–7):519–528. [PubMed: 16756576]

Ragsdale KA, Bedwell JS. Relationships between dimensional factors of psychopathy and schizotypy. Frontiers in psychology. 2013; 4 [PubMed: 23898320]

Raine A. Schizotypal and borderline features in psychopathic criminals. Personality and Individual Differences. 1992; 13(6):717–721.

Robins, LN., Tipp, J., Przybeck, T. Antisocial personality. In: Robins, LN., Regier, D., editors. Psychiatric Disorders in America. New York: Free Press; 1991. p. 258-290.

Rogers R, Jordan MJ, Harrison KS. Facets of psychopathy, Axis II traits, and behavioral dysregulation among jail detainees. Behavioral Sciences & the Law. 2007; 25:471–483.

Ronningstam E. Narcissistic Personality Disorder: A Current Review. Current Psychiatry Reports. 2010; 12:68–75. [PubMed: 20425313]

Sherry A, Henson RK. Conducting and interpreting canonical correlation analysis in personality research: A user-friendly primer. Journal of Personality Assessment. 2005; 84(1):37–48. [PubMed: 15639766]

Smith GT, McCarthy DM, Zapolski TC. On the value of homogeneous constructs for construct validation, theory testing, and the description of psychopathology. Psychological assessment. 2009; 21(3):272. [PubMed: 19719340]

Sullivan EA, Abramowitz CS, Lopez M, Kosson DS. Reliability and construct validity of the Psychopathy Checklist-Revised for Latino, European American, and African American male inmates. Psychological Assessment. 2006; 18:382. [PubMed: 17154759]

Tabachnick, BG., Fidell, LS. Using multivariate statistics. New York: Harper & Row; 2007. Thompson B. A primer on the logic and use of canonical correlation analysis. Measurement and

Evaluation in Counseling and Development. 1991; 24:80–93.

Ullrich S, Borkenau P, Marneros A. Personality disorders in offenders: Categorical versus dimensional approaches. Journal of Personality Disorders. 2001; 15:442–449. [PubMed: 11723878]

Author Manuscript

(18)

Vitacco MJ, Neumann CS, Jackson RL. Testing a four-factor model of psychopathy and its association with ethnicity, gender, intelligence, and violence. Journal of consulting and clinical psychology. 2005; 73(3):466. [PubMed: 15982144]

Vitale JE, Newman JP, Serin RC, Bolt DM. Hostile attributions in incarcerated adult male offenders: An exploration of diverse pathways. Aggressive Behavior. 2005; 31(2):99–115.

Walsh Z, Allen LC, Kosson DS. Beyond social deviance: Substance use disorders and the dimensions of psychopathy. Journal of personality disorders. 2007; 21(3):273–288. [PubMed: 17536940] Walsh Z, Swogger MT, Kosson DS. Psychopathy and instrumental violence: Facet level relationships.

Journal of Personality Disorders. 2009; 3:416.

Walters GD, Knight RA, Grann M, Dahle KP. Incremental validity of the Psychopathy Checklist facet scores: Predicting release outcome in six samples. Journal of Abnormal Psychology. 2008; 117:396. [PubMed: 18489215]

Warren JI, Burnette M. The multifaceted construct of Psychopathy: Association with APD, clinical, and criminal characteristics among male and female inmates. International Journal of Forensic Mental Health. 2013; 12:265–273.

Weizmann-Henelius G, Viemerö V, Eronen M. Psychopathy in violent female offenders in Finland. Psychopathology. 2004; 37:213–221. [PubMed: 15353887]

Wright AG, Thomas KM, Hopwood CJ, Markon KE, Pincus AL, Krueger RF. The hierarchical structure of DSM-5 pathological personality traits. Journal of abnormal psychology. 2012; 121(4): 951. [PubMed: 22448740]

World Health Organization. The ICD-10 classification of mental and behavioural disorders: Diagnostic criteria for research. Geneva: World Health Organization; 1993.

Author Manuscript

(19)

Highlights

• Shared variance between psychopathy and PDs explained by two pairs of functions

• First function explained by elevations of the interpersonal and antisocial facets

• This function was related to Paranoid, Narcissistic, Histrionic, and Antisocial PDs

• Second function explained by high antisocial facet and low interpersonal facet

• This function was related to Borderline PD

Author Manuscript

(20)

Figure 1.

Graphical depiction of the first two pairs of functions produced by canonical correlation analysis (see Results section). Each pair of functions represents the latent correlation between the four psychopathy facets and the 10 Personality Disorders (PDs). Arrows indicate the contribution of variables in each set to the corresponding synthetic function (only structure coefficients (rs) greater than |.45| are reported).

Author Manuscript

(21)

Author Manuscript

Table 1

Studies Examining Zero-Order Correlations between PCL/PCL–R Psychopathy and Personality Disorders

Study Participants Results

Hart and Hare (1989) Male forensic psychiatric patients from Canada

Positive correlations with Factor 1:

Antisocial (M), Narcissistic (M), Histrionic PDs (M) Negative correlations with Factor 1:

Avoidant (M), Dependent PDs (M) Positive correlation with Factor 2: Antisocial PD (L)

Hildebrand and de Ruiter (2004)

Male forensic psychiatric patients from the Netherlands

Positive correlations with Factor 1:

Antisocial (M), Histrionic (M), Paranoid (M), Narcissistic PDs (L) Positive correlations with Factor 2:

Borderline (M), Paranoid (M), Antisocial PDs (L) Blackburn (2007) Male forensic

psychiatric patients from England and Scotland

Positive correlations with Interpersonal facet: Antisocial (M), Histrionic (L), Narcissistic PDs (L) Negative correlations with Interpersonal facet: Avoidant PD (M)

Positive correlations with Affective facet:

Borderline PD (S), Antisocial PD (M), Histrionic (L), Narcissistic PDs (L) Positive correlations with Lifestyle facet:

Borderline (S), Histrionic (S), Narcissistic (S), Paranoid (S), Antisocial PDs (L) Negative correlations with Lifestyle facet:

Obsessive–Compulsive (S), Avoidant PDs (S) Positive correlations with Antisocial facet: Paranoid (S), Borderline (M), Antisocial PDs (L) Negative correlation with Antisocial facet: Obsessive–Compulsive PD (S)

Rogers, Jordan, Harrison (2007)

Male and female offenders at a county jail in the U.S.

Positive correlation with Interpersonal facet: Schizotypal (M), Narcissistic PDs (L) Positive correlation with Affective facet: Narcissistic PD (M)

Positive correlation with Lifestyle facet: Antisocial (M), Histrionic (M), Avoidant PDs (L) Positive correlation with Antisocial facet: Avoidant (M), Antisocial (L), Borderline PDs (L) Warren and Burnett

(2013)

Male and female prisoners from the U.S.

Positive correlations with Interpersonal facet:

Borderline (S), Paranoid (S), Schizotypal (S), Antisocial (M), Histrionic (M), Narcissistic PDs (L)

Negative correlation with Interpersonal facet: Avoidant PD (S)

Positive correlations with Affective facet:

Paranoid (S), Schizotypal (S), Antisocial (M), Narcissistic (M), Schizoid PDs (M) Negative correlations with Affective facet:

Avoidant (S), Dependent PDs (S) Positive correlations with Lifestyle facet:

Schizotypal (S), Schizoid (S), Dependent (S), Borderline (M), Narcissistic (M), Histrionic (M), Paranoid (M), Antisocial PDs (L)

Positive correlations with Antisocial facet:

Borderline (S), Histrionic (S), Paranoid (S), Schizotypal (S), Schizoid (S), Narcissistic (M), Antisocial PDs (L)

Negative correlation with Antisocial facet: Obsessive–Compulsive PD (S)

(22)

Author Manuscript

Table 2

Mean, Standard Deviation (SD) and Zero-Order Correlations of Personality Disorder Traits and Psychopathy Facets (N = 76).

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 1. Paranoid PD – .42 *** .47 *** .55 *** .57 *** .21 .67 *** .41 *** .40 *** .52 *** .25 * .13 − .13 .19 .47 *** 2. Schizoid PD – .60 *** .44 *** .28 * .11 .28 * .38 ** .32 ** .30 ** .07 − .09 .00 .13 .20 3. Schizotypal PD – .60 *** .37 ** .12 .39 *** .40 *** .52 *** .42 *** .25 * .07 .08 .29 * .29 * 4. Borderline PD – .44 *** .38 ** .56 *** .38 ** .43 ** .33 ** .27 * − .01 − .01 .32 ** .45 *** 5. Histrionic PD – .13 .59 *** .24 * .27 * .39 ** .39 ** .38 ** .10 .26 * .44 *** 6. Antisocial PD – .24 * .07 .05 .03 .35 ** .20 † .18 .20 .34 ** 7. Narcissistic PD – .26 * .23 * .37 ** .32 ** .31 ** − .07 .18 .39 ** 8. Avoidant PD – .56 *** .55 *** .11 − .01 − .04 .16 .18 9. Dependent PD – .56 *** .15 − .004 − .06 .31 ** .16 10. Obsessive-Compulsive PD – .14 .15 − .17 .22 .11 11. PCL-R total – .83 *** .74 *** .74 *** 64 *** 12. PCL-R Interpersonal – .57 *** .49 *** .35 ** 13. PCL-R Affective – .46 *** .28 * 14. PCL-R Lifestyle – .34 ** 15. PCL-R Antisocial – M 5.38 2.63 3.74 4.78 3.93 9.82 5.76 1.74 1.39 3.53 25.95 4.68 5.04 6.91 7.18 SD 3.64 1.97 3.21 2.85 2.57 2.27 4.09 2.09 1.80 2.89 7.03 2.37 1.97 1.95 2.13 * p < .05; ** p < .01; *** p < .001. Note.

(23)

Author Manuscript

Table 3

Canonical Correlation Analysis of Dimensionally Assessed Personality Disorders and Psychopathy Facets (N = 76).

Variable Function 1 Function 2 h 2 (%) Coef. rs Coef. rs Personality Disorders Paranoid − .24 − .73 51.84 .52 .43 18.49 70.33 Schizoid .22 − .13 1.69 .12 .41 16.81 18.50 Schizotypal − .14 − .35 12.25 − .14 .39 15.21 27.46 Histrionic − .41 − .77 59.29 − .11 .12 1.44 60.73 Borderline .30 − .43 18.49 .98 .76 57.76 76.25 Narcissistic − .50 − .85 72.25 − .73 .08 0.64 72.89 Antisocial − .33 − .47 22.09 .02 .29 8.41 30.50 Avoidant − .04 − .26 6.76 .00 .25 6.25 13.01 Dependent .08 − .20 4.00 .25 .43 18.49 22.49 Obsessive–Compulsive − .09 − .43 18.49 − .28 .05 0.25 18.74 43.69 38.15 Psychopathy facets Affective .59 − .06 0.36 .06 − .06 0.36 0.72 Interpersonal − .81 − .70 49.00 − 1.00 − .46 21.16 70.16 Lifestyle − .03 − .35 12.25 .55 .33 10.89 23.14 Antisocial − .61 − .74 54.76 .68 .54 29.16 83.92 Note. Coef.

= standardized canonical function coefficient.

rs

= structure coefficients.

= squared structure coefficient.

h

2 = communality coefficient.

rs

greater than |.45| and

h

2 greater than 45