Strengths and limitations of actuarial prediction of criminal reoffence in a German prison sample: a comparative study of LSI-R, HCR-20 and PCL-R

Unlike many other countries, for many years, Germany disregarded structured instruments for assessing the risk of criminal reoffence. However, this negative attitude now seems to be gradually changing. An increasing number of contributions regarding structured instruments have been published in the German literature in the last years, and some instruments have already found their way into practice. However, studies that systematically examine the applicability of the mostly Anglo-American instruments to German criminals are still lacking. Therefore, the major objective of the current study was to test some internationally established procedures in a larger unselected sample from the German penal system. The following were included in the study: the Level of Service Inventory - Revised (LSI-R), the HCR-20 Scheme, and the Psychopathy Checklist - Revised (PCL-R). On the whole, the instruments proved to be applicable to German criminals with only a few adaptations to the German situation, and they achieved a predictive accuracy comparable to the values reported internationally. However, there were only minor differences in the predictive performance between the measures. Moreover, some basic limitations became apparent. Firstly, we found quite high percentages of criminals with medium scores and a correspondingly ambiguous prognosis. Furthermore, the predictive accuracy seemed to be dependent on demographic, criminological and psychopathological characteristics of the offenders. Finally, the instruments appeared to only partially utilize the empirical store of knowledge available regarding factors influencing the recidivism of criminals, since even a simple predictive model that only added a few further aspects besides the tested instruments (e. g. treatment yes or no) achieved systematically better predictions than the instruments alone. Altogether, the tested measures turned out to be useful instruments for risk assessments and may be conducive for a more systemized practice. However, due to the limitations inherent, they should be seen as a complement to a careful and clinically informed appraisal and not a substitute.