Gender disparity in the prediction of recidivism: the accuracy of the LSI-R modified

During the last 50 years, the rate at which females enter the correctional system has increased exponentially. Despite this influx, risk assessment instruments remain geared toward male offenders. The Level of Service Inventory-Revised (LSI-R) is considered by some to be one of the most predictive and comprehensive risk instruments, but critics assert that this instrument neglects risk factors salient for female offenders. This study examined whether modifying the LSI-R to assess gender responsive variables (i.e., victimization, economic marginality, and "gendered" substance abuse) would result in an improved assessment of recidivism risk over the original LSI-R. Participants were 37 male and 26 female offenders incarcerated at community corrections centers and county jails in a southeastern state. The study found that the inclusion of all the gender responsive crime variables did not increase the predictive accuracy of the LSI-R. However, the victimization domain performed better than the other gender responsive variables in increasing the predictive accuracy of the LSI-R, while not impacting the predictive accuracy for male offenders. Furthermore, the victimization domain accounted for a significant amount of variability in the rearrest status, of both male and female offenders, above and beyond that predicted by the LSI-R risk score. Implications regarding the assessment of dynamic victimization factors in risk evaluation practices are discussed.