The effectiveness of Corrections' rehabilitation interventions with Māori

Dr Peter Johnston - Director Research and Analysis, Department of Corrections

Dr Peter Johnston has been with the Department of Corrections for nearly 30 years. He started with the Psychological Service in Christchurch, as one of three psychologists who set up the first special treatment unit, Kia Marama, at Rolleston Prison in 1989.  He then moved to the (then) Prison Service, where he was involved in setting up prisoner assessment centres and designing an end-to-end case management system. In his current role since 2004, he leads a team of ten staff who undertake research and evaluation, and conduct in-depth analysis of criminal justice data to shed light on trends and developments in the offender population, measure the impacts of rehabilitation, and to support strategy and policy initiatives.

Introduction

The Department delivers a reasonably wide range of programmes and interventions to enable offenders to lead law-abiding lives. Research shows that significant reductions in reconviction and re‑imprisonment can be achieved when well-designed interventions are delivered to appropriately-selected offenders.

The Department measures the impact of its rehabilitation and re-integration activities through the “Rehabilitation Quotient” (“RQ”). RQ gauges the extent to which re-offending is reduced, by comparing rates of reconviction and re‑imprisonment amongst offenders who received a rehabilitative intervention, with the rates recorded amongst offenders who have an equivalent risk of re‑offending, but who had no exposure to the particular intervention.

RQ allows for the fact that a significant number of offenders in any given cohort will have completed multiple rehabilitative and re-integrative interventions. Factors unrelated to the particular intervention under scrutiny are “controlled for” — in other words, all other known factors which could have an influence on outcomes are held constant. This allows us to see how effective a specific intervention is in reducing re-offending amongst those who participated in and/or completed it.

RQ scores are reported as a number between zero and ~20, equating to the percentage-point changes in rates of either re-imprisonment or reconviction. Another term for these figures is the programme’s “effect size”. The RQ score is based on the percentage-point difference between treated and untreated offender groups; thus, an RQ re-imprisonment score of 10 might indicate, for example, that the rate of re-imprisonment amongst untreated offenders was 35 percent, and the corresponding rate for the programme “graduates” was 25 percent. By international standards, effect sizes of 10 percentage points are considered an excellent outcome[1].

Limitations of RQ results are recognised. While the RQ methodology involves matching of offenders in terms of a wide range of risk-relevant characteristics (e.g. age, gender, ethnicity, sentence length, sentence type, RoC*RoI scores[2]), as well as by sentence start or end dates, there is no random assignment of offenders to “treatment” and “untreated” (comparison) groups. Consequently, there is potential for some selection bias to influence scores.

Research generally shows that RQ scores of between 7 and 15 percent are attainable when good quality services are delivered in a targeted manner to appropriately selected offenders. As noted, effect sizes of 10 percentage point reductions can be considered very good outcomes.

RQ results for Māori participants

The Department has sought to understand the differential impact of its programmes on Māori participants.  It is not feasible to produce RQ scores for Māori across all rehabilitation and reintegration activities, mainly due to too-small sample sizes; further, the demands of the task means that analysis of results by ethnicity cannot be undertaken every year. However the Department has accumulated evidence to show it is achieving statistically significant reductions in re-offending for Māori across certain interventions.  The main ones are as follows:

  • In each of the last six years effect sizes of between 8 and 13 percent have been recorded for prisoners completing the Special Treatment Unit Rehabilitation programmes (STURP;  measured as RQ for re-imprisonment in the 12 months following release). Given that a majority of participants in each annual programme cohort are Māori, these results indicate high levels of positive impact with Māori participants.  Subsequent analysis of the 2016/17 results confirmed that, when analysed separately, Māori performed as well, or slightly better, than non-Māori participants.
  • In 2015 an 8 percent effect size was recorded for Māori prisoners who completed the Te Tirohanga programme (measured as RQ for re-imprisonment in the 12 months following release). This compares to a 4.1 percent RQ re‑imprisonment effect size for all (Māori and non-Māori) prisoners who completed the Te Tirohanga programme.
  • An 8.5 percent effect size was recorded for Māori prisoners who completed the “Out of Gate” programme (measured as RQ for re-imprisonment in the 12 months following release); this compares to a 6.2 percent RQ re‑imprisonment effect size for all (Māori and non-Māori) prisoners who completed the “Out of Gate” programme.
  • When the records of all participants in prison-based employment training are combined, RQ analysis indicates that effect sizes for Māori prisoners (7.4 percent reconviction, 5.8 percent reimprisonment) are superior to the aggregated results for all participants.
  • In 2015 an RQ effect size (reimprisonment) of 7.6 percent was recorded for the programme Mauri Tu Pai, delivered within the Te Tirohanga units, and in which the participants are almost exclusively Māori. Over the last ten years, lesser effect sizes, between 3.1 and 5.7 percent (re imprisonment) have been recorded for individual years.

The Department also has evidence to show it is achieving good effect sizes for Māori across other interventions, even though these are just below the threshold for statistical significance (i.e., it is likely the programme is having a positive effect). Many programmes that ultimately produce good effect sizes “start small” so these programmes may be relatively new, or they may be run for reasons not solely focused on reducing re-offending, such as building motivation for change.

  • Over the last four years, RQ effects sizes of between 2.7 and 6.4 percent (reconviction) have been achieved by the programme Kowhiritanga, delivered within the women’s prisons and in which the participants are predominantly Māori.
  • A 6.7 percent effect size was recorded for Māori prisoners who completed the three-month Drug Treatment Unit (“DTU”) programme (measured as RQ for re‑imprisonment in the 12 months following release). This compares to a 5.3 percent RQ re‑imprisonment effect size for all prisoners (Māori and non-Māori) who completed the three-month DTU.
  • A 6.6 percent effect size was recorded for Māori prisoners who completed the six-month DTU) programme (measured as RQ for re‑conviction in the 12 months following release). This compares to a 4.8 percent RQ re‑conviction effect size for all prisoners (Māori and non-Māori) who completed the six-month DTU.
  • A 5.5 percent effect size was recorded for Māori prisoners who completed the Medium Intensity Rehabilitation Programme (“MIRP”) (measured as RQ for re-imprisonment in the 12 months following release). This compares to a 5.2 percent RQ re‑imprisonment effect size for all prisoners (Māori and non-Māori) who completed the MIRP.
  • A 3.9 percent effect size was recorded for Māori offenders serving community sentences who completed the Domestic Violence Programme (measured as RQ for re‑conviction in the 12 months following release). This compares to a 4.2 percent RQ re‑conviction effect size for all offenders (Māori and non-Māori) serving community sentences who complete the Domestic Violence Programme.

It should be kept in mind that RQ results can vary from year to year, by programme type, and by participant sub-group.  In general it appears safe to conclude, however, that Māori participants respond as well as non-Māori to the Department’s mainstream rehabilitative interventions.

Responses of gang members to rehabilitation

According to Corrections data, it has been determined that 70 percent of gang members in prison are Māori. A separate project involving RQ-style analysis was conducted in 2013, focusing on outcomes for gang members who participated in Departmental rehabilitation programmes.  This revealed the following:

  • Gang-affiliated prisoners were participating in rehabilitation in numbers only slightly below what would be expected given the proportion of the prisoner population they comprised
  • Participation rates varied from programme to programme, with (for example) few gang members participating in sex offender special treatment units (STUs), but significant numbers in the Māori Focus Units (MFUs) and Māori Therapeutic Programmes (MTPs).
  • Across a range of programmes, RQ effect sizes (i.e. percentage-point differences in rate of reimprisonment between gang members who participated, and risk-matched gang members who didn’t) were as large, and sometimes larger, than the effect sizes found for non-gang member participants; this clearly indicates that gang members were obtained some benefits from participating in rehabilitation.
  • However, raw rates of reconviction and reimprisonment amongst gang members who participated in programmes were nevertheless invariably higher (and often much higher) than the rates of non-gang participants.  Reimprisonment rates for gang participants were even higher than the rate for non-gang non-participants.

Summary

The Department remains committed to addressing the relatively high rates of re-offending amongst Māori offenders. This means delivering a range of programmes, both in prisons and the community, which address the key drivers of offending behaviour in ways that are culturally sensitive to all participants. Many of these programmes have been developed with strong Māori input and all have relevant cultural components.  In addition, there is a range of culturally-based services which promote a greater sense of identity and connection to whānau and iwi.  In the main, the Department has been achieving very promising gains though these programmes and services, but the quest to improve the magnitude of these gains, in terms of reduced re-offending, continues.

[1] Results of these analyses are presented in the Department’s annual reports – see https://www.corrections.govt.nz/resources/strategic_reports/annual-reports.html
[2] Risk of (Re-)Conviction*Risk of (Re-)Imprisonment, the Department’s actuarial re-offending risk measure.