Prevalence and prediction of institutional misconduct and gradual release violations in socio-therapeutic institutions

The prediction of institutional misconduct and gradual release violations is essential for the treatment planning in prisons and in most cases a precondition for decisions about gradual release and other incarceration related efforts. However, currently existing risk assessment tools have not yet been examined in the German speaking language area to determine whether they are also suitable for predicting institutional misconduct and gradual release violations. In this study, three actuarial risk assessment tools (Static-99, SVG-5, and OGRS 3) were examined in this context based on 129 offenders, who were released from the socio-therapeutic institution in Ludwigshafen (Rhineland-Palatinate, Germany) between 2013 and 2018. In addition, two internal checldists were included in the study to assess their predictive validity. All three actuarial instruments and the institutional checklist about the risk of flight showed good predictive accuracy for the prediction of institutional misconduct, with the OGRS 3 (AUC = .77) achieving the highest effect sizes. The results of the OGRS 3 and SVG-5 were also promising for the prediction of gradual release violations, with the OGRS 3 (AUC = .77) again obtaining the highest predictive accuracy. However, especially for the prediction of gradual release violations, the results also showed limitations of the available methods, which have to be taken into consideration when they are used in applied risk assessment settings.