Do Static-99R and PCL-R Scores Predict Risk Opinion Equally Across Evaluators in SVP Civil Commitment Evaluations?

Jurisdictional legislation attempts to provide the public protection from sexual offenders. Sexually Violent Predator (SVP) statutes allow for civil commitment of sexual offenders considered high risk for perpetrating additional sexual crimes upon completion of their punitive sentence. SVP evaluations integrate multiple considerations and although structured risk assessment measures decrease the incidence of evaluator differences, research suggests evaluator differences continue to impact clinical opinions. When multiple sources of information contribute to the evaluation, practitioners may place varying levels of importance on any given piece of data. This raises the question: What data do evaluators rely on most strongly to form their ultimate opinions? The current study examined whether scores from two measures used in the course of SVP evaluations, the Static-99R and the Psychopathy Checklist-Revised (PCL-R), predicted ultimate risk opinions and determination of behavioral abnormality equally across evaluators. Data was extracted from the evaluation records of 393 incarcerated men convicted of sexual offences who were evaluated by one of three evaluators for civil commitment in the state of Texas. It was expected that Static-99R scores would predict risk opinion equally across evaluators. With regard to PCL-R factors and facets, it was expected that Factor 2 scores, Facet 4 in particular, would predict risk opinion equally across evaluators, while Factor 1 scores might differentially predict risk. Results indicated that PCL-R Facet 4 scores did predict risk opinion equally across evaluators, as expected. All other relationships between measure score and risk opinion differed across evaluators. With regard to behavioral abnormality, it was expected that Static-99R scores would be differentially predictive across evaluators, but that PCL-R scores would be equally predictive. Results showed that PCL-R Facet 4 scores did equally predict behavioral abnormality opinion across evaluators. Static-99R scores predicted opinion about behavioral abnormality as well, which was not expected. All other relationships between measure scores and behavioral abnormality differed by evaluator, also contrary to hypothesis. Potential contributing factors to these varied results are discussed.