Predictive Validity of the Static-99R in Romania

Abstract

Empirical actuarial measures are significantly more accurate than unstructured clinical judgments in predicting recidivism risk of incarcerated populations across different cultures. In a retrospective study, we examined the predictive validity of the Static-99R, the actuarial measure most used around the world for assessing the recidivism risk of persons who committed a sexual offense (PCS), in a sample of 518 male PCS in Romania. The results showed a good predictive validity for any kind of recidivism (area under the curve [AUC] = .70) and a slightly lower predictive accuracy for violent (AUC = .69) and sexual recidivism (AUC = .67). Effect sizes and the predictive validity of the single items of the Static-99R have also been examined. In addition, the predicted rates of reconviction have been compared with the observed rates. The Romanian version of the Static-99R has been shown as a valid, useful tool to predict sexual, violent, and general recidivism for Romanian PCS.