By Emma Ziersch, Shawn Sowerbutts, Yilma Woldgabreal, Sophie Ransom

The Violence Risk Scale (VRS) is a tool designed to assess and predict risk of future offending, inform decisions around therapeutic intervention and measure changes in violence risk as a result of treatment. This study investigated the discriminative and predictive validity of the VRS for Australian Aboriginal and non-Indigenous males convicted of violent offending in multiple jurisdictions.

The Violence Risk Scale (VRS) is a risk assessment tool designed to assess and predict risk of future offending, inform decisions around therapeutic intervention and measure changes in violence risk as a result of treatment. While the tool has been used extensively both internationally and in Australia, its applicability to our Australian population is unclear.

This study investigated the discriminative and predictive validity of the VRS for Australian Aboriginal and non-Indigenous males convicted of violent offending in multiple jurisdictions. The VRS total score had moderate discriminative accuracy for violent reoffending at five-year follow-up. However, Aboriginal males were significantly more likely to be categorized as high risk, and additional discrimination measures revealed variation in performance between Aboriginal and non-Indigenous males. Implications of the findings for correctional practice and recommendations to reduce bias in the assessment of Aboriginal offenders are discussed.

 Research Report no. 34. Canberra: Australian Institute of Criminology. 2025. 66p

By Alex Severson,  Elise Ferguson

This study evaluates the predictive validity of the Public Safety Assessment (PSA) in Bernalillo County, New Mexico, using data from 22,387 felony cases between July 2017 and June 2023. The PSA, which generates scores predicting defendants' likelihood of failure to appear (FTA), new criminal activity (NCA), and new violent criminal activity (NVCA) during pretrial release, demonstrated "fair" to "good" predictive validity with AUC scores ranging from 0.58-0.69, aligning with results from our initial validation study from June 2021. While the PSA showed similar predictive performance across racial groups for FTA and NCA outcomes, the NVCA flag performed poorly for Black defendants, with predictive validity no better than chance. We also found that age-related risk factors embedded in the NCA and NVCA scales had limited predictive value. We documented significant increases in FTA rates during and after COVID-19, likely due to policy changes that increased court appearances through mandatory status hearings. Despite higher base rates of NCA and NVCA for male defendants, the PSA demonstrated similar predictive validity across genders. Analysis of NCA by charge severity revealed that even high-risk defendants rarely committed serious felonies while on pretrial release. Our findings suggest opportunities to optimize the PSA locally by adjusting how age-related factors are weighted and reconsidering the use of the NVCA flag for Black defendants.

Albuquerque: University of New Mexico, INSTITUTE FOR SOCIAL RESEARCH CENTER FOR APPLIED RESEARCH AND ANALYSIS , 2024. 48p.