A base data set was extracted which included all offenders potentially eligible for front-end Home Detention during the 2004/05 financial year - that is, all offenders sentenced during that period to a term of imprisonment of two years or less. This particular cohort was selected so that, at time of the analysis, all such sentences would have been completed, which in turn ensured that it was possible to examine any subsequent breaches of the order.
Around 85% of all sentenced prisoners annually have sentences of two years or less; thus the data extract produced a sample of 7606 instances of offenders sentenced to imprisonment 1. Data was extracted on each offender with respect to the following variables:
- age (at time of sentencing);
- risk score (RoC*RoI) 2;
- the most serious offence to which the conviction and sentence related;
- sentence length;
- gang affiliation (Y/N) 3;
- any previous conviction history for violent offences;
- any previous conviction history of administrative offences involving non-compliance with sentences or orders;
- (if Leave to Apply was granted) whether the offender did or did not apply for Home Detention;
- offender applications and application withdrawals for Home Detention; and
- Parole Board decisions and reasoning regarding Home Detention.
Back-end home detainees were excluded for several reasons: the majority of offenders on Home Detention (77%) are there as a result of the “front-end” process. Further, the considerations that apply to the former order are quite distinct to front-end Home Detention, and this (“back-end”) pathway to Home Detention will be significantly altered under the proposed Criminal Justice Reform Bill.
As noted above, the data extract included Parole Board decisions to either approve or decline each Home Detention application; these records included brief narrative summaries of the reasons given by the Parole Board for each decision. A categorisation system was then developed to thematically classify reasons.
From the same data source it was also possible to identify those who had obtained Leave to Apply, but who then chose not to make an application, or who withdrew their application after initially filing it. Unfortunately no data was available on the offenders’ reasons for such decisions.
In cases where there were multiple applications and/or hearings regarding the single custodial episode, only the final application or hearing was selected.
The main data set was subjected to sophisticated analysis by an independent statistician 4 to quantify the relative contribution made by each of the identified factors to judicial decision-making; the specific analysis conducted sought to demonstrate the extent to which factors other than ethnicity were influential. A separate report giving the results of this analysis is attached.
In addition to the analysis of the above data, interviews were conducted with three members of the Parole Board, to gain further insights into the decision-making process, and to better understand how the Parole Act is currently being applied with respect to Home Detention. Interviews were also conducted with CPS personnel with experience and expertise in processing Home Detention applications.
In the sample of offenders examined, it was found that Pacific offenders experienced rates of being granted Leave to Apply, and of approvals at Parole Board hearings, only slightly lower to those for New Zealand European offenders 5. The comparable rates for Maori offenders on both counts however were considerably lower than these other two groups. Further, inspection of the Home Detention data indicates that Pacific offenders are not under-represented in that muster. Given the small margin observed between European and Pacific offenders, further analysis of possible reasons for such differences was not attempted; the report therefore focuses solely on analysis of differences between New Zealand European and Maori offenders.
1. It was possible for offenders in the sample to be represented more than once, by having multiple episodes of "custody starts" within the 12 month period: in fact, 524 offenders were included in the sample more than once, accounting for 1136 records. A single offender even managed to commence eight discrete custody episodes in this 12-month period. Approximately twice as many Maori (319) as Europeans (157) had more than one episode of custody start within the year
2. 'RoC*RoI" stands for "Risk of Re-Conviction, Risk of Re-Imprisonment", a statistically-based formula which produces estimates of an individual offender’s likelihood of being reconvicted or re-imprisoned. Scores are derived primarily from the conviction and sentencing history of the offender.
3. Gang status is recorded as a result of information from Police, offender self-report or as a result of prison staff identifying membership.
4. The statistical analysis was undertaken by John Horwood, Senior Research Fellow at the Christchurch School of Medicine, and principal statistician for the internationally acclaimed Christchurch Health and Development Study.
5. Leave to apply granted to 39% of Europeans and 37% of Pacific offenders; 19% of Europeans, and 17% of Pacific offenders subsequently approved for Home Detention