Exploring Mechanisms of Change in the Rehabilitation of High-Risk Offenders
The success or failure of many different types of treatment is often measured by one type of outcome. For example, treatment for substance abuse might be judged to have failed if a patient “goes on a bender” some time after completing the programme. The same is true for offender rehabilitation. Treatment success or failure is usually determined by whether or not an offender is reconvicted of a new offence in a specified follow-up period. We know from the literature that offender rehabilitation can have modest but significant effects on reducing recidivism. Yet we know little about what brings about these reductions (i.e., how the treatment worked). This thesis explores possible mechanisms of change in offender rehabilitation. I propose that although a reduction in recidivism is an important long-term outcome of treatment, there are a number of additional outcomes that have the potential to explain not only if but how treatment works and why it is unsuccessful in leading to a reduction in reoffending for some offenders. Study 1 is a typical outcome evaluation of New Zealand’s rehabilitation programmes for high-risk male offenders: the High Risk Special Treatment Units (HRSTUs). I compared the recidivism rates of a sample of HRSTU completers with a comparison sample of high-risk offenders who had not completed the programme (a between-subjects design). I found that relative to the comparison group, treatment completers had significantly lower rates of four different indices of recidivism, varying in severity. The remainder of the thesis explored possible mechanisms of change within the HRSTU sample (a within-subjects design). Study 2 examined immediate outcomes of treatment, which I defined as within-treatment change on dynamic risk factors. I found that offenders made significant change on the Violence Risk Scale during treatment, but there was no significant relationship between treatment change and recidivism. Studies 3 and 4 examined intermediate outcomes of treatment, which I defined as barriers (risk factors) and facilitators (protective factors) that influence the process of offender re-entry. Study 3 validated an instrument designed to measure these factors: the Dynamic Risk Assessment for Offender Re-entry (DRAOR). I found that the tool had good convergent validity and reliably predicted recidivism above a static risk estimate. Study 4 used the newly validated DRAOR to test an explanation for the lack of a direct relationship between treatment change and recidivism. I tested whether treatment change had an indirect relationship with recidivism through its influence on the re-entry process. I found that treatment change was related to a number of re-entry outcomes; however, only two models could be tested for mediation because the re-entry outcomes themselves lacked predictive ability. Nevertheless, findings from Study 4 suggest the re-entry process is an area worthy of further investigation. Taken together, the findings from this thesis highlight the importance of considering alternative treatment outcomes in addition to whether or not a programme leads to a reduction in long-term recidivism outcomes. Answering the question of how treatment works requires an exploration into possible mechanisms of change. This thesis was only a preliminary investigation into such mechanisms; however, the findings have both practical and theoretical implications for the way we conceptualise how treatment programmes work. Developing a greater understanding of mechanisms of change in offender rehabilitation has the potential to lead to the design and delivery of more effective programmes.