Wong, Stephen C. P.

The predictive validity of the Violence Risk Scale (VRS) has been demonstrated in a number of institutionalised offender samples but not in a community sample. The present study assessed prospectively, in a seven-year follow-up, the validity of the VRS ratings in predicting recidivism in a sample of 60 federal offenders residing in the community after they were released under conditional orders. Six recidivism outcomes were used: any and violent reconvictions, days until any and violent reconvictions, frequency and severity of any new reconvictions. Pearson correlations, Receiver Operating Characteristic and survival analyses were used, among others, to assess the predictive validity of VRS ratings. The VRS ratings significantly predicted all six recidivism outcomes. The results also suggest that release decision makers correctly identified offenders with lower than average risk and recidivism rate for release. However, for release decision making, it is suggested that more attention should be paid to the results of risk assessments using validated structured risk assessment tools.