Assessing women’s risk of recidivism: An investigation into the predictive validity of the Dynamic Risk Assessment for Offender Re-entry (DRAOR) with matched samples of community-sentenced women and men

Although men and women share risk factors for offending, some scholars claim these factors operate differentially by gender and that certain proposed women-specific risk factors are neglected in the existing gender-neutral risk assessment tools. The present research evaluated one such gender-neutral risk assessment tool used by New Zealand Department of Corrections: The Dynamic Risk Assessment for Offender Re-entry (DRAOR; Serin, 2007; Serin, Mailloux, & Wilson, 2012). The research was comparative and examined the predictive validity of the DRAOR for breaches of sentence and criminal reconvictions in matched samples of New Zealand women and men who had served community supervision sentences. Cox regression and AUC analyses showed the initial DRAOR had mixed predictive validity and the proximal DRAOR comparative predictive validity across gender. Additionally, the proximal DRAOR assessment consistently outperformed the initial DRAOR in the prediction of reconvictions for both women and men. Further, offenders made significant change on the DRAOR between two assessment points and overall the change made on the DRAOR was significantly related to reconvictions for women and men. For both samples, the RoC*RoI did not predict breach reconvictions; however, the proximal DRAOR TS provided incremental predictive validity above the RoC*RoI for criminal reconvictions. To conclude, the research supports the continued use of the DRAOR as a risk prediction tool with community-sentenced women and men and thus supports gender neutrality. Further, the research supports the dynamic nature of the DRAOR and highlighted the importance of updating dynamic risk assessments. Additionally, the research recommends that change made on a dynamic risk assessment tool over time be considered useful for predictive purposes for women and men alike.