Psychopathy & Sexual Recidivism

October 9, 2011
Some researchers have found significant relationships between psychopathy and sexual recidivism (Hanson & Harris, 2000; Hemphill, Hare, & Wong, 1998; Hildebrand, de Ruiter, & de Vogel, 2004; Salekin et al., 1996) whereas other researchers have found competing results (Barbaree, Seto, Langton, & Peacock, 2001; Olver & Wong, 2006). Barbaree et al. (2001) found that the PCL-R predicted general and serious recidivism but not sexual recidivism. Similarly, Olver and Wong (2006) found no relationship between psychopathy and sexual recidivism. It is important to note that sexual recidivism base rates can be altered depending on the researcher’s definition of which offenses are included in sexual recidivism and in the length of follow-up time allowed (Olver & Wong, 2006). As a result, the differing definitions may have contributed to the opposing findings stated above. Olver and Wong defined sexual recidivism as any charge or conviction for a sexual offense.


As described earlier, Salekin et al. (1996) examined two studies addressing sexual sadism and deviant sexual arousal, producing effect sizes of .77, .58, and .47 with a mean of .61. Meaning, these authors found that the PCL-R Factor 2 scores may be able to predict sexual sadism.


Hanson and Harris (2000) researched static and dynamic predictors of sexual offense recidivism. Their study consisted of a total of 409 participants: 208 recidivists and 201 non-recidivists. Hanson and Harris found that recidivists were more likely to meet the criteria for Antisocial Personality Disorder (64% v. 49%, p < .002) and even more so of psychopathy (21% v. 8%, p < .001).


In contrast, Olver and Wong (2006) examined psychopathy, sexual deviance, and recidivism among a sample of sex offenders within a 9-year follow-up period. Overall, Olver and Wang found that 33% of the sample was charged or convicted of a new sexual offense and 59% received any nonsexual conviction. Moreover, Olver and Wang found that the PCL-R total (cut off = 25), Factor 1, and Factor 2 scores were weak predictors of sexual recidivism (Area Under the Curve, AUC = .61, .60, and .57, respectively). These authors noted that when evidenced, Factor 1 appeared to contribute to the prediction of sexual recidivism rather than Factor 2. In contrast, Olver and Wang discovered larger and significant correlations between PCL-R total scores and all measures of nonsexual recidivism (AUC = .73). They consistently found that Factor 2 rather than Factor 1 made the largest contribution to nonsexual recidivism (AUC = .71 and .61, respectively).

Return to Blog Main Page