Assessing the utility and validity of adolescent violence risk assessment approaches in an Australian young offender population

There is a paucity of literature examining the predictive accuracy of widely used juvenile violence risk assessment instruments in an Australian youth justice context. Additionally, less is known about the ability of such instruments to extend across gender and ethnic minority groups within Australian young offender populations. Demonstrating the efficacy of youth risk measures enables the identification of high risk offenders, subsequently guiding rehabilitation and intervention strategies. The study investigated the predictive validity and cross cultural applicability of three violence risk inventories developed in Canada in a sample of 213 young male and female offenders in custody in Victoria, Australia. The Structured Assessment of Violence Risk in Youth (SAVRY), the Youth Level of Service/Case Management Inventory (YLS/CMI) and the Psychopathy Checklist: Youth Version (PCL: YV) were employed to ascertain the level of risk of the participants. Recidivistic offenses were recorded for up to 18 months for participants who were released from custody during the study. The validity of the risk instruments was determined by the association between risk level, total scores and domain scores, and instances of general and violent recidivism. For the overall sample and for male participants, all three instruments moderately predicted both violent and general re-offense. Across ethnic subgroups, the SAVRY ably predicted re-offense for English Speaking Background participants, though demonstrated no utility for the Culturally and Linguistically Diverse (CALD) group. Furthermore, the SAVRY was able to establish validity across particular domains for female and Indigenous (IND) participants, though a number of the findings did not achieve statistical significance due to their lower overall representation in Australian custodial settings. SAVRY item scores were also investigated for their individual interaction with the re-offense outcome and their comparative prevalence across gender and ethnicity. The results of the studies indicate that the SAVRY is a useful instrument in identifying salient risk factors for violence in an Australian young offender population. Second, the SAVRY, YLS/CMI and the PCL: YV are capable of predicting general and violent recidivism with moderate accuracy for a typical adolescent custodial sample in Victoria, Australia. Third, the SAVRY instrument demonstrated encouraging utility for female and Indigenous participants though further research is required in Australian conditions with larger samples to determine generalizability. Lastly, the SAVRY was unable to predict re-offense for the multi-ethnic CALD category. Limitations of the findings and implications for future research are discussed