The aim of this study was to compare the long-term predictive validity of the PCL-R for offenders with severe mental disorders, substance abuse and personality disorders and for offenders without mental disorders. The sample consisted of 262 perpetrators who were assessed for their criminal responsibilities. The PCL-R was assessed retrospectively from file data. Participants were prospectively followed-up for an average observation period of 58.6 months (range -138 months), with the first entry into the official criminal records of the National Conviction Registry serving as the endpoint. PCL-R scores were significantly different between the three diagnostic groups. The highest PCL-R scores were found in offenders with personality disorders and/or substance abuse. In all three groups the PCL-R predicted violent, but not non-violent reoffenses with moderate predictive validity. The PCL-R is therefore a moderately strong predictor for violent reoffenses in offenders with mental disorders, personality disorders and/or substance abuse and for those offenders without mental disorders, but failed to predict non-violent criminal recidivism.

The validity of the VRAG was replicated with a German sample. The VRAG yielded good predictive accuracy, despite differences in sample and outcome variables compared with its original sample.