| Recent efforts to expand control and scrutiny of sexual offenders have led to the increased utilization of actuarial prediction instruments. A number of these instruments now utilized throughout the United States have been validated on development samples, and a growing number of studies have been done to cross-validate them or to compare their predictive capabilities. In addition, there is almost no research designed to evaluate the utility of the instruments to predict specific types of offenses. This study compares the predictive capabilities of three sex offender risk assessment tools: the California Actuarial Risk Assessment Tables (CARAT; Schiller & Marques, 1999), The Rapid Risk Assessment for Sexual Offense Recidivism (RRASOR; Hanson, 1997), and the Static 99; Hanson & Thornton, 2000). Utilizing a case-controlled design, the files of all male sex offenders released from Minnesota prisons in 1988 and 1990 (N = 311) were scored and coded on the items on the three assessment tools and on additional variables identified by the author of this study. The three measures were found to predict reoffense approximately 25% better than chance. None of the measures, however, appeared to predict reoffense better than the others. In addition, none of the tools were able to discriminate reoffenders by category. |
