Psychopathy & Recidivism

October 6, 2011
More recent debate and research has focused on the use of specific assessments of offender personality traits believed to correlate with offender recidivism, in particular the PCL-R and its derivatives. Salekin, Rogers, and Sewell (1996) reviewed 18 studies, yielding 29 effect sizes, that investigated the relationship between the PCL/ PCL-R and violent and nonviolent recidivism. These authors divided the studies into the following groups: 1) predictions of violent recidivism and institutional violence, 2) predictions of general recidivism, and 3) predictions of sexual sadism and deviant sexual arousal.


The effect sizes for violent recidivism and institutional violence ranged from .42 to 1.92 with a mean effect size of .79 (Salekin et al., 1996). These authors found substantially lower effect sizes for general recidivism. In this category, the effect sizes ranged from .24 to .93 with a mean effect size of .55. Lastly, these authors examined two studies addressing the last category, sexual sadism and deviant sexual arousal. This produces effect sizes of .77, .58, and .47 with a mean of .61. That is, the PCL and PCL-R were very effective at predicting violent recidivism and institutional violence, less effective at predicting general recidivism, and the PCL-R Factor 2 scores may be able to predict sexual sadism. These authors concluded that psychopathy is associated with for criminal (mean d = .55) and violent behavior (mean d = .79) in male offenders. Furthermore, the psychopathic traits predict violence in those diagnosed with a serious mental disorder (mean d = .53) in addition to sexual sadism, deviant sexual arousal, and sexual recidivism in inmates (mean d = .61).


Walters (2003) conducted a meta-analysis examining the psychopathy checklist factor score’s ability to predict institutional maladjustment and recidivism. This author included 42 studies yielding 50 different effect sizes. The results of the meta-analysis indicated that Factor 2 ( mean rw = .29) was significantly more predictive of recidivism than Factor 1 (mean rw = .18). In addition, the total PCL-R score was not more predictive of institutional maladjustment and recidivism than Factor 2.


In another study, Leistico, Salekin, DeCoster, and Rogers (2008) conducted a large-scale meta-analysis relating the Hare measures of psychopathy to institutional misconduct. They integrated the effect sizes of 95 studies (N=15,826). All mean effect sizes were significantly greater than zero, thus, higher PCL Total, Factor 1, and Factor 2 scores were associated with increased recidivism or institutional misconduct. The weighted mean d for PCL Factor 2 was .60, followed by .55 for PCL Total, and .38 for PCL Factor 1 In summary, their meta-analysis revealed that impulsive and antisocial behavioral traits of psychopathy had a stronger relation with institutional misconduct and had better predictive ability at longer follow up periods than the affective and interpersonal traits. Thus, using the PCL-R as a measure of institutional misconduct and post-release outcomes is moderately supported.

Return to Blog Main Page