The Sex Offender Risk Appraisal Guide

Abstract

In this chapter, we describe the development, validation, replication, and recommended forensic application of the actuarial Sex Offender Risk Appraisal Guide (SORAG), especially its application in the United States. Because the SORAG is designed to predict violent rather than sexual recidivism specifically, we outline why we regard rapsheet violent recidivism as the outcome closest to subsequent sexually violent behavior as specified in U.S. civil commitment laws for sex offenders. We also present preliminary data on a mean 20-year follow-up of 750 male sex offenders from our earlier studies. The violent failure rate of these men over the entire follow-up was 58 % and the area under the ROC curve for SORAG for the prediction of violent failure was 0.73 (95 % CI = 0.69–0.77). Survival analyses suggested that the final violent failure rate of our sample would be 63 % if all non-recidivists had 28 years of opportunity. The SORAG yielded high predictive accuracy over follow-up times ranging from 6 months to over 40 years, and it also predicted severity of violent recidivism and time until first violent recidivism. We also discuss whether the SORAG or other actuarial tools based on static factors should be altered due to "dynamic" risk factors, whether there is empirical support for clinical overrides of actuarial scores, and whether scores should be reduced as an offender ages. We conclude there is no evidence to support alteration of the actuarial risk score for any of these reasons.