Can the quality of high-risk violent prisoners' release plans predict recidivism following intensive rehabilitation? A comparison with risk assessment instruments

Abstract

Even when motivated to desist from crime, many high-risk offenders fail quickly following release from prison. One cause may be a lack of preparation for release. Recent research with child-sex offenders has demonstrated that men who avoided reconvictions not only had better plans made for life on parole prior to their release, but plan quality also added significant incremental validity to the prediction of recidivism after controlling for both static and dynamic measures of risk. This study examined release planning in high-risk violent prisoners in an intensive cognitive-behavioural rehabilitation programme. We compared the predictive validity of plan quality with three well-validated risk assessment instruments. Men who were reconvicted had significantly higher scores on all three risk instruments and significantly poorer plans, but plan quality did not significantly improve prediction when risk was controlled. Plan quality was also significantly poorer in men who were reimprisoned, and did significantly improve prediction over and above each risk prediction instrument. Findings suggest that higher quality release plans may protect offenders from being quickly reimprisoned, despite high levels of assessed criminal risk, and that enhancing resources put into release planning may lead to improved parole outcomes.

Keywords:

• reintegration
• reconviction
• reimprisonment
• violent offenders
• risk assessment

Acknowledgements

This research was partially funded by the New Zealand Department of Corrections Strategy, Policy and Planning, and data collection was supported by the Departmental Psychological Service. The first author was partially supported by a New Zealand government Summer Research Scholarship.

Notes

1. In New Zealand, parole is formally defined as the conditional release of an offender who has been sentenced to two or more years in prison. Its aim is to ease the transition back into the community from prison (NZ Parole Board Website http://www.paroleboard.govt.nz/utility/faq/parole_faqs.html). However, the term parole in this article also is used more broadly, to refer to the immediate period following incarceration, regardless of attached conditions.

2. We used the VRS total score prior to treatment rather than the post-treatment score to facilitate comparisons with the PCL:SV and RoC*RoI, both of which are scored before treatment, and also to reduce potential overlap with the plan quality measure; several VRS item ratings would be based on information also used to score pre-release plan quality.

3. Because the sample included some men who did not complete the treatment programme, t-tests were repeated, excluding these men. The pattern of results was unchanged, so they were retained in order to maximise sample size.

4. The reconviction AUC for release plan quality was nonsignificant (.65, p = .07)

5. However, when the regression was performed the other way around, by entering plan quality in Block 1 and then PCL:SV in Block 2, the model is not significantly improved by the addition of the PCL:SV compared to plan quality alone.

6. Neither the PCL:SV nor plan quality were individually significant predictors when combined, though the combined model was significant.