An Analysis of Decision Making in District Prisons Boards

Abstract

The Criminal Justice Act 1985 substantially expanded the system of parole release in New Zealand. At the time of this study, approximately 90% of offenders serving custodial sentences were eligible to be considered for early release by one of the country's 17 District Prisons Boards. A substantial body of literature has investigated aspects of parole, but has typically adopted a narrow focus upon the prediction of parolee reoffending. The aim of the present study was to examine the system of District Prisons Board parole release from a broader perspective, encompassing a full description of boards' operations and procedures, a formal evaluation of the group decision making process, and evaluation of performance on a traditional measure - prediction of reoffending. In respect of the second of these objectives - the evaluation of group decision making process - a preliminary psychometric investigation revealed the selected measurement instrument to be substantially flawed. The results of these analyses have therefore been included as an appendix. With regard to boards' operations and procedures, results showed that board members lacked any clear conception of an overarching framework guiding their operations, that members were broadly oriented towards reducing reoffending, but that a strong rehabilitative ethic found them regarding parole as providing an opportunity to reduce risk of reoffending through the provision of release programmes rather than through the identification and prolonged detention of high risk inmates. Examination of decision making criteria showed members to over-report the number of factors commonly considered. A statistical model based upon seven biographical variables was able to accurately predict 77% of release decisions made by boards. The proportion of inmates released on parole was found to vary widely between boards, and these substantial variations were found to be robust when controlling for inmates' custodial security level. A 30 month follow-up found 77% of all inmates to have been reconvicted and 45% to have been returned to prison during the period. No difference was found in the reconviction rates of inmates denied parole and released at 2/3 of sentence on remission, and inmates granted parole and released at an earlier date. However, remission release inmates tended to be reconvicted earlier than parolees and, when reconvicted, were more likely to receive a further sentence of imprisonment. Ten biographical factors were found to differentiate reoffenders and non-reoffenders and, when combined in a prediction model, knowledge of these factors allowed approximately 80% of reoffenders to be correctly predicted. These findings are discussed and their implications for District Prisons Board policy outlined.