Proefschrift

2 58 CHAPTER 2 If dynamic risk factors are predominantly correlates of psychological propensities, causally related to sexual (re)offending, then treatment of these dynamic risk factors would still, at best, “treat the symptoms” and not the causes. This might be considered a less effective treatment approach and therefore could provide a basis for the small effect sizes we found for change scores. Nevertheless, positive change on dynamic risk does indicate that the risk level of a particular man with a history of sexual offenses is reduced, and this finding warrants a continued focus on the role of change scores in future research. We could not carry out moderator tests for several of our main analyses because of a lack of a sufficient number of studies in each category. However, of the moderator analyses that were possible, results suggest that the impact of publication bias was negligible. Nevertheless, some other moderator variables did show a significant impact on effect sizes. For the overall predictive properties for sexual recidivism of dynamic risk assessment instruments we found that original (vs. replication) studies, earlier published studies, studies with a shorter average follow-up interval, and studies using both institution and community samples, were associated with larger effect sizes. For violent (including sexual) recidivism, the opposite was found for average follow-up intervals (shorter follow-up intervals were associated with smaller effect sizes), and studies with a larger sample size had smaller effect sizes. We found no significant moderator effects in the analyses focusing on incremental validity, and we were not able to perform moderator analyses for change scores because of a lack of a sufficient number of studies in each category. The present study is the first meta-analysis to specifically focus on the predictive properties of dynamic risk assessment instruments developed for adult men with a history of sexual offenses and adds, we believe, valuable new knowledge on the assessment of recidivism risk. The meta-analysis was based on both published and unpublished studies from various countries and settings that used a range of risk assessment instruments. Nevertheless, some limitations should be acknowledged. For example, our meta-analysis was based on a modest number of unique study samples. Especially the analyses using change scores are based on a relatively small number of studies. Also, the fact that most samples came from studies conducted in Europe and North America limits generalizability of the findings to other, non-western adult sex offender populations. Despite these limitations, we believe the findings of this metaanalysis are promising and hopefully will inspire and encourage researchers to design and conduct further studies on dynamic risk assessment (instruments). The findings that dynamic risk assessment instruments significantly predict recidivism, add additional predictive value to static risk assessment, and that change scores significantly predict recidivism, supports the conclusion that dynamic risk assessment instruments designed for use in adult men with a history of sexual offenses can be a useful aid in optimizing forensic treatment in a way that is consistent with

RkJQdWJsaXNoZXIy MjY0ODMw