Sex offenders with “deadly combination” of psychopathy and deviant sexual interests are not more likely to reoffend - CONCEPT Professional Training

Paige B. Harris, Sam Houston State University
Marcus T. Boccaccini, Sam Houston State University
Amanda K. Rice, Sam Houston State University

Offenders with high levels of both psychopathy and deviant sexual interests are often described as being more prone to recidivate than other sexual offenders, and many forensic evaluators report considering this psychopathy and sexual deviance interaction when coming to conclusions about sex offender risk. However, empirical support for the interaction comes from studies using sexual deviance measures that are rarely used in the field. We examined the ability of Psychopathy Checklist-Revised (PCL-R) field scores and possible field measures of sexual deviance (e.g., paraphilia diagnosis, offense characteristics) to predict sexual recidivism among 687 offenders released after being evaluated for postrelease civil commitment (M follow-up = 10.5 years). PCL-R total scores and antisocial personality diagnoses were predictive of a combined category of violent or sexual recidivism, but not sexual recidivism. Paraphilia diagnoses and offense characteristics were not associated with an increased likelihood of reoffending. There was no evidence that those with high levels of both psychopathy and sexual deviance were more likely than others to reoffend. Although the psychopathy and sexual deviance interaction findings from prior studies are large and compelling, our findings highlight the need for research examining the best ways to translate those findings into routine practice.

psychopathy, PCL-R, sexual deviance, risk assessment, sexual offenders

“Offenders with high levels of both psychopathy and deviant sexual interests—the so called “deadly combination” of sex offender traits (Hare, 1999, p. 189)—are often described as more prone to recidivate than other sexual offenders (Hare, 2003; Witt & Conroy, 2008).” (p. 639)
“A recent survey of sex offender risk assessment practices revealed that the two most commonly used field measures of sexual deviance were a documented history of deviant sexual behavior (96%) and a paraphilia diagnosis (83%; Boccaccini et al., 2017). None of the psychopathy and deviance interaction studies have directly examined these commonly used field measures of sexual deviance.” (p. 639–640).
“Although standardized deviance instruments rely, to varying extents, on some of the same diagnosable and documented behaviors used by clinicians in the field, findings from standardized measures may not generalize to the less structured diagnostic and clinical judgment practices evaluators use in the field.” (p. 640).
“Evaluators in field settings have varying levels of training and experience, and there is often no oversight of their assessment and scoring practices. Research has shown that systematic evaluator differences in psychopathy measure scoring tendencies can lead to higher levels of measurement error in field scores than research scores.” (p. 640)
“There are also reasons to expect a significant amount of measurement error in some field measures of sexual deviance, particularly diagnoses. […] Although diagnoses are widely applied in practice, there is no specific training required for assigning diagnoses and almost certainly a large amount of variability in the practice of assigning them.” (p. 640)
“Although many evaluators report using information about the combined pattern of psychopathy and sexual deviance when coming to conclusions about sex offender risk, the interaction study literature is probably smaller and more variable than many evaluators suspect.” (p. 640–641)
“Our goal was to conduct the first field validity study of the sexual deviance measures that evaluators report using in the field and to examine whether offenders with high levels of deviance and psychopathy were more likely to reoffend than other offenders.” (p. 641)
“We obtained PCL-R scores, diagnoses, and information about postrelease sexual and violent offenses from 687 sexual offenders who were released from custody after being evaluated for civil commitment as sexually violent predators (SVP). […] Participants were 687 male sexual offenders who were evaluated for civil commitment as SVPs, but released (i.e., not committed) after their evaluations.” (p. 641)
“We collected information for this study from evaluator’s behavioral abnormality evaluation reports and a copy of the records that the Texas Department of Criminal Justice (TDCJ) provides to the evaluators. These records include information about index and prior offenses, prison disciplinary infractions, and prior testing conducted by TDCJ staff (e.g., Static-99, Personality Assessment Inventory).” (p. 642)
Measures used: Psychopathy Checklist Revised (PCL-R), Diagnosis, Documented history of deviant sexual behavior, Screening Scale for Pedophilic Interests (SSPI). Postrelease arrest data was used to assess recidivism rates.
“Using 10 possible field measures of sexual deviance, we found little evidence that offenders with higher levels of deviance were more likely to reoffend than others.” (p. 645)
“There was also no evidence that offenders with high levels of both psychopathy and sexual deviance were more likely to reoffend than other offenders.” (p. 646)
“There are at least three possible explanations for our deviance measure findings. First, it may be that our field measures of sexual deviance do not measure sexual deviance, or do not measure sexual deviance adequately enough for there to be an interaction effect. […] A second possible explanation for our findings is the questionable field reliability of our psychopathy and deviance measures. […] A third possible explanation is that there is something unique about our sample.” (p. 646–648)
“It seems more likely that evaluators rely on paraphilia diagnoses and offense characteristics when coming to conclusions about deviance. […] However, none of these deviance indicators were predictors of recidivism in this study.” (p. 648)
“PCL-R scores and antisocial personality disorder diagnoses were predictive of the combined category of violent or sexual recidivism, but not sexual recidivism. […] Although some have argued that this combined category of violent and sexual arrests may be a better indicator of true sexual recidivism than sexual arrests alone (Rice, Harris, Lang, & Cormier, 2006), the small negative effect for post-release sexual arrests in our sample argues against using antisocial personality disorder diagnoses for predicting sexual recidivism.” (p. 648)

“This study found that the types of offender and offense characteristics that field evaluators report using as indicators of sexual deviance were not predictive of postrelease sexual offending among a large sample of sexual offenders who underwent risk assessments before release from prison. In this field study, there was no evidence that offenders with high levels of both psychopathic traits and deviant sexual interests—the so called “deadly combination” of sex offender traits—were more likely to reoffend than other sexual offenders.” (p. 639)
“Although evaluators report using paraphilia diagnoses and documented incidents of deviant sexual behavior as their primary field measures of sexual deviance, we found that no field measure produced the same type of interaction effect documented in prior studies. Thus, at this point, there does not appear to be sufficient empirical support for using the combination of PCL-R scores and these field measures for coming to conclusions about offender risk, at least in the context of SVP evaluations. Evaluators who wish to base their risk assessment practices on documented empirical support should look to studies reporting significant interaction effects (e.g., Harris et al., 2003; Olver & Wong, 2006; Seto et al., 2004), and the deviance measures used in those studies, such as plethysmography or the Violence Risk Scale: Sex Offender version (Wong, Olver, Nicholaichuk, & Gordon, 2004).” (p. 649)

“The average age at the time of the evaluation was 42.82 years (SD = 11.44). The number of sexual offense victims ranged from one to eight (M = 2.54, SD = 1.08). Although offenders eligible for SVP civil commitment must have been convicted of at least two contact sexual offenses and be serving a sentence for a sex offense at the time of evaluation (Texas Health & Safety Code, Title 11, Chapter 841, 2000), the qualifying sexual offenses can be against the same victim.” (p. 641)
“We did not include offenders who had been civilly committed because of the intensive monitoring and supervision within the SVP program. […] There is no doubt that the exclusion of committed offenders affected the study sample, but the extent to which their exclusion might explain the difference between our findings and prior studies is less clear. […] Although there would have been more variability in our sample if we had been able to study recidivism among the committed offenders, there was no evidence that our sample differed dramatically from the samples used in other psychopathy and sexual deviance studies. […] We also have no information about participation in sexual offender treatment or postrelease supervision, factors which may help explain the low base rate of recidivism in this study.” (p. 648)
“Although we did not find any evidence of an interaction between antisocial personality diagnoses and sexual deviance for predicting recidivism, this is clearly an area in need of more research.” (p. 648)
“Although our findings add to a growing body of research suggesting weaker reliability and validity in field settings, they also highlight possible areas for growth. Nonfield studies show us that instruments and assessment practices can attain desired levels of reliability and validity, and it is possible for field practices to improve. […] Documenting the current performance of field practices allows us to better understand where we are underperforming, helps us to identify areas in need of improvement, and provides a baseline for future studies aimed at improving in field performance.” (p. 649)

As always, please join the discussion below if you have thoughts or comments to add!