Psychopathy & Female Offenders

October 17, 2011
There is a paucity of research on the validity of using the PCL-R or its derivatives with female offenders (Salekin, Rogers, & Sewell, 1997). These authors elaborated that pronounced gender differences for psychopathy and APD have been observed, despite the limited research. Some of the gender differences include later onset of psychopathy and APD for females than males; females exhibit less overt aggression than males; and females have higher rates of comorbid psychiatric disorders than males. These authors concluded that much critical research needed regarding the construct of psychopathy and its relation to females. 


Kennealy, Hicks, and Patrick (2007) review the available literature and suggest that females exhibit a delayed onset and lower prevalence of early behavioral problems, as well as lower prevalence of violent and aggressive behavioral problems. Factor 1 items, especially interpersonal facet items, have been reported to be less prominent in females. Thus, these gender differences could be the result of a gender bias present in the clinical interview, file information, or in the PCL-R item criteria themselves (Kennealy et al., 2007). Another possibility is that females with psychopathy do exhibit interpersonal deficits, but they differ in kind from those in males, and thus are not well assessed by current PCL criteria.


Strand and Belfrage (2005) examined gender differences in psychopathy in a Swedish offender sample. Their sample was comprised of 129 female psychiatric patients, forensic psychiatric evaluees, and criminal offenders, which was compared to a reference group of 499 male in similar settings. Strand and Belfrage found that 16% (N = 20) of the female group and 25% (N = 123) of the male group were psychopathic based on PCL:SV total scores. Using item response theory analysis, Strand and Belfrage found that females were more impulsive and had worse behavioral control which may explained by the fact that more females than males were diagnosed with Borderline Personality Disorder, whereas, more males were diagnosed with Antisocial Personality Disorder. However, these authors noted that both of these personality disorders are related to psychopathy. Strand and Belfrage concluded that the prevalence in their study was comparable to that found in other studies regardless of methodology and instruments being used.


Hare explained that there is evidence that the PCL-R is a valid measure to use with female offenders (Hare, 2004). Kennealy et al. (2007) examined the validity of the PCL-R in female offenders by examining the discriminant relations of the PCL-R factors with criterion measures. The participants consisted of 226 female offenders from a correctional institution (Kennealy et al., 2007). The participants were rated on the PCL-R, original Cleckley criteria, DSM-IV-TR APD criteria, criminal history, and social background variables. These authors found that the PCL-R was highly correlated with Cleckley’s criteria (r=.48 to.66, p<.01). Additionally, the results indicated that the PCL-R had large positive bivariate correlations with child and adult APD symptoms. Similarly, the PCL-R was significantly correlated with criminal history and social background variables (i.e. years of education, maternal occupational status, maternal criminality, and paternal criminality). These results revealed significant convergence and divergence across measures supporting the construct of psychopathy in female offenders. Therefore, the use of the PCL-R was deemed a valid measure of psychopathy for female offenders. 


Similarly, Hart et al. (2003) reported that correlations between PCL:SV Total scores and sex were significant and negative, with a weighted mean correlation of -.33. This suggests that there is a significant, although small, sex difference in the prevalence of psychopathic symptoms, with males experiencing more symptoms than females.


Nicholls, Ogloff, Brink, and Spidel (2005) reviewed the clinical utility of the PCL-R in diverse samples of females. These authors observed that psychopathy research in community samples is still in its infancy. The majority of the research has been derived from samples of undergraduates who do not clearly represent the general population. Additionally, few prospective studies exist that have researched the predictive validity of psychopathy with future criminality, aggression, and violence in non-forensic samples. That being said, these authors concluded that the results of their review consistently suggested that about 1% of the general population, and fewer females than males, are likely to meet diagnostic criteria for psychopathy (female PCL-R mean = 7.0, SD = 5.8; male PCL-R mean  = 11.5, SD = 7.2).


Furthermore, Nicholls et al. (2005) discovered that mean PCL-R and PCL:SV scores were higher for male forensic psychiatric patients (PCL-R: M = 7.8, SD = 3.9; PCL:SV: M = 11, SD = 4.54) than female forensic psychiatric patients (PCL-R: M = 6.1, SD = 2.9; PCL:SV: M = 10.48, SD = 4.78). That is, males had higher psychopathy scores than females. Finally, the total scores appeared to have utility in prediction of aggression, violence, and offending for both male and female forensic patients. These authors concluded that these measures may perform as well, and possibly better, with female than male forensic psychiatric patients. Further research is needed to clarify the nature of these findings.


Warren et al. (2003) assessed the concurrent validity of the PCL-R and the HCR-20, a risk assessment measure, to better understand the risk factors for violence and criminality in female offenders. Their sample consisted of 132 female inmates from a maximum-security prison. Warren et al. found that there was a high degree of correlation between the PCL-R and the HCR-20 (with PCL-R scores) total scores (r=.80, p=.001)[1].


Warren et al. (2003) found differing PCL-R and HCR-20 results depending on the type of crime that was committed, specifically regarding murder, property, and minor crime cateogories. Their research revealed that female offenders who had been convicted of murder scored significantly lower on the PCL-R (M=19.18) and HCR-20 (M=17.5) than those female offenders who had only been charged with murder. Moreover, female offenders who had convictions of property crimes scored higher on both the PCL-R (M=25.1) and HCR-20 (M=23) than female offenders who did not have property crime convictions. A similar pattern was found in the minor crimes categories, with females having convictions scoring significantly higher on the PCL-R and HCR-20. These authors did not find significant differences on the PCL-R and HCR-20 for violent, potentially violent, sex, and drug crimes. These authors concluded that psychopathic female offenders appeared to be significantly at risk for future violence and therefore, further validates the use of the PCL-R with the female offender population.


The relationship between psychopathy and recidivism among female jail detainees was examined by Salekin, Rogers, Ustad, and Sewell (1998).  Recidivism data on a sample of 78 female inmates were examined at one-year post-release using the PCL-R, Personality Disorder Examination, and PAI Antisocial Features and Aggression scales. These authors discovered that the prevalence of psychopathy in females based on the PCL-R was 12.9% (n=10). Further, 14.1% (n=11) of the participants were classified as psychopathic on the PAI Antisocial Features scale with a T-score of at least 80.


The recidivism rate for the entire sample of females was 41%. Ironically, Salekin et al. (1998) found that female psychopaths were less likely to recidivate in a 14-month follow-up period than male psychopaths. Fifty percent of the 10 females who scored above 29 on the PCL-R were rearrested for a criminal offense, whereas 50% of those who had a PCL-R score above 29 did not recidivate. In comparison, this rate is 12.6% lower than the average rate of recidivism for male psychopaths in previous studies. Over 90% of female recidivists were not identified by the PAI ANT scale. Further, 18% of the nonrecidivists were classified as psychopathic by the ANT scale. According to these authors this may have been due to misclassification by the PAI ANT scale, however, it is important to keep in mind that not all psychopaths recidivate. Although the PCL-R was somewhat better with classification than the PAI, the results were not impressive. The results indicate that the construct of psychopathy as measured by the PCL-R and its ability to predict recidivism generalizes moderately, but perhaps not completely, to female offenders. Overall, these authors concluded that that the above measures of psychopathy do not perform as well with females as they do with males.


Next, using a nearly identical dataset as the above study, Salekin et al. (1997) assessed the construct validity of psychopathy in female offenders. Their sample consisted of 103 female offenders in a county jail. The PCL-R and PAI were administered to all participants. These authors found that 15% of female offenders were designated to be psychopathic when using the PCL-R cut-off score established by Hare (>29) compared to 30% of males. Furthermore, these authors demonstrated strong convergent validity of psychopathy with similar personality disorders, as evidenced by high correlations with the Antisocial (.68), Paranoid (.61), and Borderline (.60) clinical scales of the PAI. These authors found minimal evidence of convergent and discriminant validity of the PAI subscales and PCL-R factor scores and suggested that the 2 Factor model of psychopathy may not be applicable to females in general. Factor 1 is more appropriately characterized by lack of empathy, interpersonal deception, proneness to boredom, and sensation seeking. Factor 2 is more characterized by early behavioral problems, promiscuity, and adult antisocial behavior. Further, the results in the present study support the relationship between psychopathy and aggression. The PAI Aggression scale was highly related to all three measures of the syndrome (PAI-ANT and A-PDE scales (rs > .70) and moderately related to PCL-R (r = .49). Overall, these researchers found that the construct of psychopathy was applicable to female offenders but at lower prevalence rates than males. These authors concluded that the construct of psychopathy appeared applicable to female offenders. However, some refinements in the construct are indicated for females.


The aforementioned studies suggest that the PCL-R and its derivatives can be reliable and valid when used with a female offender population. However, the literature is limited and there are clear sex differences such as prevalence rates, PCL-R total scores, and recidivism rates. What these differences mean as applied to the construct and use of psychopathy measurements in women is unclear. Furthermore, Nicholls et al. (2005) explained that gender differences play an important role with many other psychological illnesses and personality disorders. These authors noted that it would be useful to have studies comparing the characteristics and backgrounds of males and females who demonstrate psychopathic traits. Additionally, Nicholls et al. emphasized the need to study the construct of psychopathy in other community samples. A review of the literature indicates a deficit regarding sex offenders, especially within a community based evaluation and treatment program. As Nicholls et al. emphasized, it would be useful to have a better understanding of psychopathic male and female characteristics. Furthermore, research suggests that sexual offenders with more psychopathic characteristics and higher in deviant sexual arousal may be at a greater risk of recidivating than sexual offenders with lower scores on these indices (Serin, Mailloux, & Malcolm, 2001). The current study plans to follow up some of these suggestions to further clarify the nature of psychopathy among male and female sexual offenders.



Return to Blog Main Page