Psychopathy in women: Structural modeling and comorbidity
Introduction
The concept of psychopathy has elicited attention and scientific study from the medical and psychiatric communities for over two hundred years. In 1801 Philip Pinel identified a pathological condition of the emotions, “mania sans delirè,” which he described as being characterized by emotional lability and social instability. The same condition acquired alternative names by French, English, and German writers including moral insanity (Prichard, 1835), delinquente nato (Lombroso, 1876), psychopathic inferiority (Koch, 1891), psychopathic personalities (Kraepelin, 1904), sociopathy (Partridge, 1930), and semantic dementia (Cleckley, 1941). Schneider (1923) identified two types of psychopathic individuals: that is, the Gemütsamer psychopath or the smug, arrogant psychopath who primarily causes suffering to others, and the Geltungsbedürftig or the needy, demanding psychopath who experiences internal suffering from their psychic abnormality (Herpertz & Sass, 2000).
Since that time, the etiology of this and other personality disorders (PDs) has been explored from a variety of perspectives with a common consensus developing which recognizes an interaction of influences including genetic predispositions (DiLalla, Gottesman, & Carey, 2000), psychophysiological processes Dolan, 1999, Lapiere et al., 1995, Raine et al., 1994, developmental influences Cloninger et al., 1975, Marshall & Cooke, 1999, and situational conditions (Loeber & Stouthamer-Loeber, 1986). The predictive and structural integrity of the construct has also attracted vigorous research attention since the creation of the Psychopathy Checklist (PCL) by Hare (1980), the Psychopathy Checklist-Revised (PCL-R) in 1991, and Psychopathy Checklist: Screening Version (PCL-R: SV) by Hart, Cox, and Hare (1995). Based on Cleckley's (1941) articulate clinical description of the condition, these empirical instruments have been found to demonstrate a robust ability to predict reoffense rates (Hare, McPherson, & Forth, 1988), institutional adjustment (Gacano, Meloy, Sheppard, Speth, & Roske, 1995), treatment response Ogloff et al., 1990, Rice et al., 1992, and community violence among prisoners (Hare & McPherson, 1984), forensic patients (Rice & Harris, 1992), and civilly committed psychiatric patients Douglas et al., 1999, Monahan et al., 2001. The strength and consistency of these findings have resulted in an enthusiastic response by professionals performing risk assessments for violence, institutional security, and community management within both the psychiatric and correctional communities.
In the present study, we examine the gender issues that emerge when exploring the performance and structure of psychopathy in women. Using a sample of 138 incarcerated female inmates, we examine seven structural models and the comorbidity of psychopathy with the 10 PDs. Our goal is to determine similarities and differences in the construct as it applies to women and men, and to contribute to the emergent research that is seeking to understand the personality dimensions of this widely recognized construct.
Over the past 15 years, research efforts to explore psychopathy's underlying structure and to demonstrate across samples its ability to predict violence and recidivism have involved primarily, if not exclusively, male samples of prison inmates and forensic patients. The structural inquiries began with the original analyses by Harpur, Hakstian, and Hare (1988), who used a split-half cross validation method to explore data collected across six samples of incarcerated and forensically hospitalized males (N=1119). Based on the analyses of congruence coefficients, they concluded that psychopathy was composed of two factors that encompassed 20 of the 22 items that were contained in the instrument at that time. The first factor, which came to be known as the Interpersonal/Affective factor, was found to include core personality traits including superficiality, habitual lying and manipulation; callousness; and a lack of affect, guilt, and remorse. The second factor termed Social Deviance was characterized by a chronically unstable and antisocial lifestyle.
Since the formulation of this model, it has been replicated nationally and internationally Cooke, 1995, Pham et al., 1998, Côte & Hodgins, 1989, Hobson & Shine, 1998 and has been applied to adolescents Murrie & Cornell, 2000, Brandt et al., 1997, community samples (Livesley, Jackson, & Schroeder, 1992), and to groups obtained from twin registries (Livesley & Schroeder, 1991). Monahan et al.'s (2001) large-scale study examined the risk for violence among discharged civilly committed psychiatric patients and found that scores on the Social Deviance dimension predicted violence among this group upon return to the community, even when 15 covariates were entered into the predictive equation (including criminal and violence history, substance abuse and diagnoses, other PDs, anger, and demographic characteristics). Structural modeling of these data, however, suggested that the two-factor model provided only an adequate fit to the data, leading Skeem and Mulvey (2001) to conclude that the marginal goodness-of-fit measures may have been due to the low base rate of psychopathy in this sample and/or structural differences in the PCL: SV when applied to civil psychiatric samples.
Cooke and Michie (2001) have been the most explicit in their study of the structure of psychopathy and began their inquiry based on the observation that the original Harpur et al. (1988) analyses were flawed by the misrepresentations of the congruence coefficients. Using a combination of eight Canadian and two American samples (N=2067), they applied confirmatory factor analyses (CFAs) to the data and further concluded that the two-factor model did not meet acceptable standards of congruence using various goodness-of-fit measures.
Cooke and Michie (2001) have subsequently sought to develop a more precise model of psychopathy using theoretical premises based on personality theory in general and psychopathy in particular combined with newer statistical techniques designed to explore the dimensional structure of the construct. The theoretical inquiry led them to define three congruent aspects of personality: (a) affective, interpersonal, and behavioral domains; (b) a hierarchical structure of personality constructs; and (c) the assumption of continuity between normal personality traits and PDs (Widiger, 1998). The statistical analyses used the direct oblimin criterion for obtaining an oblique rotation within exploratory factor analysis (EFA), combined with the application of item response theory (IRT) to create item “testlets” and cluster analysis to determine the optimal agglomerate of the paired items. These analyses converged on a three-factor model composed of 13 items, which included six testlets. The first factor, termed Arrogant and Deceitful Interpersonal Style, was made up of two testlets including glibness and grandiose sense of self-worth combined with pathological lying and conning/manipulative. The second factor, labeled Deficient Affective Experience, was composed of two testlets, shallow affect and callous/lack of empathy, and lack of remorse/guilt and failure to accept responsibility. The third factor, referred to as Impulsive and Irresponsible Behavioral Style, was made up of two testlets: the first, a need for stimulation/proneness to boredom, impulsivity, and irresponsibility; and the second, parasitic lifestyle and lack of realistic long-term goals. This model was found to result in a fit that was superior to that achieved with the two-factor model and appeared congruent with a hierarchical model subsumed within a coherent construct.
Cross validation of the three-factor model has included replication on a Scottish sample of 247 prisoners; analyses of data using the screening version of the PCL-R, the PCL-SV; and, an examination of convergent validity using the DSM-IV diagnostic criteria for antisocial PD. The diagnostic validation, which used 506 males and females selected from psychiatric, drug treatment, and at-risk populations (biological child of persons diagnosed with antisocial PD), found some degree of fit between three PCL-R factors, a DSM-III-R diagnosis of antisocial PD, and a ICD-10 diagnosis of dyssocial PD (World Health Organization, 1992), although the antisocial and dyssocial personality constructs failed to identify items that could be construed to represent the Deceitful Interpersonal Style factor of their proposed model (Cooke & Michie, 2001).
Most recently, Hare has begun to explore a four-factor structure of psychopathy made up of Cooke's three factors combined with a fourth factor, one that he terms Persistent and Varied Rule Breaking (Hare, personal communication). This research, which is being replicated on a large combined sample in Canada, has not yet been published but is included in the current analysis in a preliminary fashion.
Another approach to understanding the structure and meaning of psychopathy has involved the exploration of its comorbidity with the 10 PDs recognized by DSM-IV (APA, 2000). Blackburn (1998) and Coid used the PCL-R and SCID-II interviews with a sample of 18 psychopaths and 68 nonpsychopaths and found that psychopaths suffered from a variety of Axis I disorders, including alcohol abuse, dysthymia, depression, and schizophrenia. The only significant differences, however, involved a higher prevalence of drug abuse and a more frequent history of somatization disorder among the psychopaths. When comparisons were made with PD diagnoses, PCL-R scores were found to correlate positively with paranoid PD, antisocial PD, borderline PD, narcissistic PD, and passive aggressive PD, and negatively with dependent PD. Blackburn concluded that these results replicated those reported in an earlier study by Hart and Hare (1994) and support the notion of psychopathy as constituting a broad dimension of PD.
Based on Schneider's (1923) theory of psychopathy, Nedopil, Hollweg, Hartmann, and Jaser (1998) examined the relationship between psychopathy and major forms of mental illness including schizophrenia, dementia, substance dependence, and the various Axis II PDs. They derived a psychopathy estimate from a 512-item Forensic Psychiatric Documentation System (FPDS) and found little overlap with the major forms of mental illness. However, of 114 individuals thought to be psychopathic, 27% met criteria for substance dependence, 37% for minimal brain dysfunction, and 41% for PDs primarily of the dissocial, histrionic, and narcissistic types. The authors acknowledge the preliminary nature of these findings, and emphasize the need for further research to clarify the controversy between psychopathy as a distinct clinical entity or a comorbid combination of malignant character traits.
Widiger and Lynam (1998) have sought to interpret the psychopathy construct within the five-factor model (FFM) of normal personality functioning. Using the five dimensions of personality recognized in the FFM model (i.e., neuroticism, extraversion, openness, agreeableness, and conscientiousness), they have applied each of the 20 PCL-R items to this framework. Based upon this exploration, they conclude that the psychopath is a person characterized by high antagonism, low conscientiousness, and low anxiety. Clinically they suggest that persons with this constellation of characteristics will “invariably” be of immediate and substantial concern to other members of society. Theoretically, they suggest that psychopathy is a unique collection of personality traits and not a homogeneous, clinical entity.
Only recently has the psychopathy construct been explicitly applied to women and explored in terms of its structure, relevant cut-off scores, and associated traits and behaviors. The research that has been conducted has focused on university students Forth et al., 1996, Forth et al., 1993, Lilienfeld & Andrews, 1996, substance abusers Cooney et al., 1990, Rutherford et al., 1996, and incarcerated female inmates Salekin et al., 1997, Vitale & Newman, 2001. The studies examining psychopathy among incarcerated samples Loucks, 1995, Neary, 1990, Salekin et al., 1997, Strachan, 1995, Tien et al., 1993 suggest rates varying from 9% to 31%, a finding that has been both likened to and contrasted with the rate of psychopathy found in male samples.
Salekin et al. (1997) examined the construct of psychopathy among a sample of 103 detained female offenders in a local jail. Using three measures of antisocial personality, the PCL-R, Personality Assessment Inventory (PAI), and the Personality Disorder Examination (PDE), they found different rates of morbidity based upon the instrument used. When using the PCL-R only, 16% of the women were above the cut-off score for psychopathy, although 33% were elevated on the criteria for antisocial personality contained within the other two measures.
As part of this study, Salekin et al. (1997) conducted an EFA of their data to explore the relative fit of a two-factor model Cooke, 1995, Harpur et al., 1989, Hobson & Shine, 1998, Kosson et al., 1990. They found that a two-factor solution best fit their data, and that two variables uniquely cross-loaded on both factors (i.e., poor behavioral controls and impulsivity). While the item “many short-term marital relationships” did not load for either sample, the remaining two items differed among the male and female samples (i.e., “failure to accept responsibility” and “revocation of conditional release” for females as contrasted to “criminal versatility” and “promiscuous sexual behavior” for males). Integrating these findings into an alternative two-factor model, Salekin et al. describe their new Factor 1 as being characterized by a lack of empathy or guilt, interpersonal deception, proneness to boredom, and sensation seeking, whereas Factor 2 was found to contain early behavioral problems, poor behavioral controls, and adult antisocial behavior. Salekin et al. report a positive correlation between psychopathy and histrionic PD, as well as between measures of dominance, mania, and aggression on the PAI. A consistent negative correlation was found between the warmth scale of the PAI and total, Factor 1, and Factor 2 scores on the psychopathy measure, suggesting that the women scoring high on the PCL-R were lacking in warmth, empathy, and interpersonal sensitivity.
A single gender comparison study has examined the response patterns of 36 females and 36 males referred for forensic psychiatric evaluation in Sweden between 1988 and 1990 (Grann, 2000). Using pairs matched by age, ethnicity, socioeconomic background, index crime, and number of previous violent crimes, Grann (2000) performed a stepwise discriminant analysis with gender as the grouping variable and the 20 PCL-R variables as independent variables. They found psychopathy to be more common among males (31%) than females (11%), although the differences in total, Factor 1, and Factor 2 means scores were not significant. The discriminant analyses were able to differentiate between the males and females with a correct classification rate of 74% using three items-callous/lack of empathy and juvenile delinquency which were “male items,” and promiscuous sexual behavior which was a “female item.” Grann critically observed that these results could reflect actual gender differences in psychopathy among men and women, or alternatively gender biases in the types of questions asked and the information reported in psychiatric and criminal files.
A recent study by Vitale, Smith, Brinkley, and Newman (2002) contains the first large-scale examination of this concept among 528 nonpsychotic female offenders. They found a relatively low base rate of 9% among this sample, and attributed this either to a lower base rate of psychopathy among women or the inclusion of items in the PCL-R that do not adequately capture this personality construct among women. Their use of various self-report measures of psychopathology to validate the construct further demonstrated no significant relationship between scores over 30 on the PCL-R and scores on EPQ-Neuroticism scale, BAI, BDI, MPQ-PA (positive affectivity), SCL-90-GSI and estimated WAIS IQ. Significant differences were found between scores on the anxiety measure (WAS) and negative affectivity (MPQ-NA) in Caucasian women, and anxiety and estimated IQ scores among African American women. Vitale et al. suggest that these associations may have etiological significance and result in the misclassification of “ neurotic” or “secondary” psychopathy in some women, a category addressed but differentiated from “ primary” psychopaths in the early writing of Cleckley (1941). They further conclude that the relative lack of comorbidity across instruments offers discriminant validation of the description of the psychopath as an individual who is free of “ nervousness or psychoneurosis.”
To explore further the structural and diagnostic underpinnings of psychopathy specifically as it applies to women, we examined psychopathy as measured by the PCL-R among a sample of 138 women incarcerated at a maximum-security prison. Using CFA, the Hare two-factor and Cooke three-factor solutions were applied to the data to explore their relative usefulness in explaining the underlying structure of the concept in women. An analysis of the relationship of PCL-R total score and factor scores as they relate to the 10 PDs recognized by DSM-IV was also undertaken to determine the degree of comorbidity observed between psychopathy and other formulations of personality maladjustment.
Section snippets
Participants and procedures
The sample comprised 138 female felons incarcerated at a maximum-security prison for women. The demographic characteristics of the women including age and race, the length of sentence were obtained from the prison information database. Each of the 138 women had been part of an earlier study which examined DSM-IV personality diagnoses among a larger sample of 261 female inmates. The 261 inmates had been selected based upon a screening of 802 inmates, which included the Brief Symptom Inventory
Results
Table 1 summarizes scores on the various PCL-R items for the entire sample, those women who scored over the traditional cut-off of 30, and those that scored over the European cut-off of 25. Of the 138 female inmates, 17.4% of the sample had scores above Hare's cut-off of 30 for psychopathy, whereas 46.4% had scores above the European cut-off of 25. To facilitate comparison with other samples compiled within the United States, the more conservative cut-off was used for further group comparisons.
Discussion
The current findings suggest a considerable degree of similarity in the prevalence and structure of the psychopathy construct among women as compared to men. Earlier research suggests that approximately 15–25% of men incarcerated in a maximum-security prison meet diagnostic criteria for psychopathy using a 30 score cut-off (Hare, 1991). In the current sample, 17.4% of the female sample met the same criteria. This percentage supports earlier research among incarcerated females Loucks, 1995,
Acknowledgements
This research was funded by grant no. 98-CE-VX-0027 by the National Institute of Justice, Office of Justice Programs, and U.S. Department of Justice, to the University of Virginia. Points of view expressed in this document are those of the authors and do not necessarily represent the official position or policy of the U.S. Department of Justice.
References (68)
- et al.
The assessment of psychopathy in male and female noncriminals: reliability and validity
Personality and Individual Differences
(1996) A research scale for the assessment of psychopathy in criminal populations
Personality and Individual Differences
(1980)- et al.
Violent and aggressive behavior by criminal psychopaths
International Journal of Law and Psychiatry
(1984) - et al.
Selective reductions in prefrontal glucose metabolism in murderers
Biological Psychiatry
(1994) - et al.
A comparison of criminal recidivism among schizophrenic and nonschizophrenic offenders
International Journal of Law and Psychiatry
(1992) - et al.
Psychiatric symptoms, history of victimization, and violent behavior among incarcerated female felons: an American perspective
International Journal of Law and Psychiatry
(2002) Diagnostic and statistical manual of mental disorders
(2000)- et al.
The effect of sampling error on convergence, improper solutions, and goodness-of-fit indices for maximum likelihood confirmatory factor analysis
Psychometrika
(1984) Comparative fit indexes in structural models
Psychological Bulletin
(1990)Psychopathy and personality disorder: implications of interpersonal theory
Assessment of psychopathy in a population of incarcerated adolescent offenders
Psychological Assessment
The mask of insanity
The multifactorial model of disease transmission: III. Familial relationship between sociopathy and hysteria (Briquet's syndrome)
British Journal of Psychiatry
Psychopathic disturbance in the Scottish prison population: the cross-cultural generalizability of the Hare Psychopathy Checklist
Psychology, Crime, and Law
Refining the construct of psychopathy: towards a hierarchical model
Psychological Assessment
A comparison of methods for assessing sociopathy in male and female alcoholics
Journal of Studies on Alcohol
Madness beyond the threshold? Associations between personality and psychopathology
On the neurology of morals
Nature Neuroscience
Assessing risk for violence among psychiatric patients: the HCR-20 violence risk assessment scheme and the Psychopathy Checklist: Screening Version
Journal of Consulting and Clinical Psychology
The Structured Clinical Interview for DSM-III-R Personality Disorders (SCID-II): I. Description
Journal of Personality Disorders
The Structured Clinical Interview for DSM-III-R Personality Disorders (SCID-II): II. Multi-site test–retest reliability study
Journal of Personality Disorders
Precursors to psychopathic traits in a sample of male and female university students
Canadian Psychology
A clinical investigation of malingering and psychopathy in hospitalized insanity acquittees
Bulletin of the American Academy of Psychiatry and the Law
The PCL-R and gender
European Journal of Psychological Assessment
The Hare Psychopathy Checklist-Revised
Male psychopaths and their criminal careers
Journal of Consulting and Clinical Psychology
Two-factor conceptualization of psychopathy: construct validity and assessment implications
Psychological Assessment: A Journal of Clinical and Consulting Psychology
Factor structure of the Psychopathy Checklist
Journal of Consulting and Clinical Psychology
Manual for the Psychopathy Checklist—Screening Version (PCL:SV)
Psychopathy and the big five: correlations between observers' rating of normal and pathological personality
Journal of Personality Disorders
Emotional deficiency and psychopathy
Behavioral Sciences and the Law
Measurement of psychopathy in a UK prison population referred for long-term psychotherapy
British Journal of Criminology
LISREL VI: analysis of linear structural relationships by maximum likelihood, instrumental variables, and least square methods
Cited by (119)
Comparative Study on Nutrient Composition, Functional Property and Glycaemic Index of “Ogi” in Healthy Rats Prepared from Selected Cereal Grains
2022, Journal of Future FoodsCitation Excerpt :Hence, YM may be a better grain for production of Ogi samples with low glycemic index.
Psychometric psychopathy: Relationships with indices of reinforcement sensitivity theory factors
2022, Personality and Individual DifferencesAntioxidant activity and blood glucose reduction potential of Malabar Chestnut in Streptozotocin Induced Diabetic Rats
2022, Journal of Agriculture and Food ResearchTesting the utility of the psychopathy construct for predicting criminal recidivism among detained girls
2022, Journal of Criminal JusticeCitation Excerpt :We are aware of no studies that have used three APSD component scores to identify criminal justice-involved girls with a PPP. Hence, findings from our latent profile analysis must be carefully interpreted, especially since the percentage of girls with a PPP seems high in comparison to PCL-R based prevalence rates of female psychopathy (±17%) in various forensic adult samples (e.g., Lehmann & Ittel, 2012; Salekin, Rogers, & Sewell, 1997; Warren et al., 2003). Also, prior work that used three psychopathy component scores as clustering variables typically identified three (da Silva et al., 2019) or more classes (e.g., Colins, Fanti, Salekin, & Andershed, 2017; Wang et al., 2020).
Primary and secondary psychopathic traits: The role of attachment and cognitive emotion regulation strategies
2021, Personality and Individual DifferencesUnderstanding Female Offenders: Psychopathy, Criminal Behavior, Assessment, and Treatment
2021, Understanding Female Offenders: Psychopathy, Criminal Behavior, Assessment, and Treatment