Canadian Journal of Behavioural Science / Revue canadienne des sciences du comportement, Vol 58(4), Oct 2026, 297-305; doi:10.1037/cbs0000468
This study examines the presence of item bias in actuarial methods used to assess recidivism risk by focusing on their use among female offenders. Starting with the risk assessment and clinical analysis method for offenders in Quebec (ORAC-PCQ), an analysis of differential item functioning (DFI) was carried out using the Raju method integrated into a Rasch model. The goal was to determine if certain items functioned differently according to sex, which could compromise the fairness and validity of the scores produced. The results reveal the presence of DFI in several items, suggesting that the method might assess risk differently for men and women. Furthermore, the analysis highlights different risk profiles: women present more issues related to victimization, and instability in the home and at work; whereas men demonstrate a more significant history of criminality and antisocial behaviour. The results of this study will help correctional staff refine their criminological assessments while also orienting the adaptation of social reintegration programs according to gender. (PsycInfo Database Record (c) 2026 APA, all rights reserved)