Voorspelling naar inrichting

Abstract

In dit onderzoek is voor de casus van justitiële jeugdinrichtingen (JJI’s) onderzocht onder welke condities het haalbaar is om te komen tot een onvertekende (valide) en stabiele (betrouwbare) schatting van het relatieve effect van de JJI’s op recidive. Hiervoor is gekeken naar gegevens over de periode 2004 tot en met 2013. Ook is bekeken of er een standaardwerkwijze voor modellering ontwikkeld kan worden.

In the field of justice, there is often a need to compare institutions on their outflow results, especially in terms of recidivism. This can be for instance court districts, prisons or probation offices. In order to make institutions comparable in terms of recidivism, differences in the characteristics of the offenders entering the institution must be taken into account. This can only be done by developing a (statistical) model that corrects recidivism for these offender differences in intake. For the case of juvenile detention centers (JJIs), this study investigated under which conditions it is feasible to arrive at an unbiased (valid) and stable (reliable) estimate of the relative effect of the JJIs on recidivism. For this purpose, we looked at data for the period 2004 up to and including 2013. It was also examined whether a standard approach for modeling could be developed for this context. Previous WODC research on the population of hospital order (tbs) had already shown that the predictions of the method used to date were insufficiently precise.