The predictive validity of the PCL:YV and the SAVRY in a population of adolescent offenders

This study evaluated the predictive validity of two recently developed adolescent risk assessment instruments for the prediction of violent recidivism in an adolescent offender population. The Psychopathy Checklist: Youth Version (PCL:YV; Forth, Kosson, & Hare, in press) and the Structured Assessment for Violence Risk in Youth (SAVRY; Bartel, Borum, & Forth, 1999) were completed for 108 young offenders admitted for a court-ordered psychiatric assessment between 1990 and 1991. Outcome data consisted of the provincial conviction records for these young offenders between the ages of 18 and 21 years. Case outcomes were equally split into three categories of recidivism, including violent recidivism, non-violent recidivism, and no recidivism. The predictive validity of the instruments was evaluated using analyses of variance, correlations, receiver operator characteristic analyses, odds ratios, logistic regression analyses, and survival curve analyses. Results indicated that the total score on the PCL:YV was generally a stronger predictor of violent recidivism than the total score on the SAVRY, but a weaker predictor than the structured clinical ratings of risk (i.e., low, moderate, high) made by the SAVRY. Implications for adolescent violence risk assessment theory, research, and practice are discussed and arguments address the differing roles for the instruments within the youth justice system. Specifically, the PCL:YV maybe most effective as a violence risk instrument influencing shorter-term management, supervision, and/or placement issues, while the SAVRY may be most beneficial as a violence risk/needs instrument guiding intervention strategies, treatment placements, and/or program evaluations.