Conclusions

Evidence on the effectiveness of some rehabilitation programmes to reduce the likelihood of future recidivism is strong. The New Zealand programmes reviewed in the current article provide further weight to this argument. The effectiveness of both general (e.g., the Psychological Service) and specialist treatment programmes (e.g., for violent offenders, driving offenders, and child sex offenders) in New Zealand compares favourably with international reviews. Recent positive results with respect to the provision of culturally responsive and appropriate treatment to Māori offenders is especially encouraging.

Yet we have by no means reached the pinnacle of treatment effectiveness, even for our so-called most effective programmes. Collectively, the studies presently reviewed have a number of limitations. Firstly, the majority of the specialist treatments appear to only reduce the re-offending rate of the behaviour under intense focus. With the exception of the DOT programme (Bakker et al., 2000) the studies thus far reviewed that do provide data on recidivism rates for specific and general re-offending suggest that general re-offending is not yet being addressed by attendance at these specialist programmes. This indicates that while antisocial attitudes, beliefs, and behaviours specific to the offending (e.g., violence and child sex offending) may be being addressed, more pervasive or generalised antisocial behaviours are remaining either unchallenged or resistant to challenge during these specialist programmes. Addressing this shortcoming offers the opportunity to substantially broaden the effectiveness of programmes and in doing so reduce both specific and general offending behaviour.

Secondly, the issue of ‘programme stacking’ deserves comment. This relates to the provision of more than one rehabilitative programme during the course of an offender’s sentence, for example, completion of a tikanga programme, followed by Straight Thinking, and immediately prior to release, a specialist violence prevention programme. Very few international studies, and no New Zealand studies, have addressed the complex issue of programme stacking (for an exception see Marquis, Bourgon, Armstrong & Pfaff, 1996). However, given the above discussion on the specificity of treatment results (i.e. treatment appears to only impact on the targeted offending behaviour and not to have a generalised effect to other types of offences), the provision of multiple types of programming – especially if targeted at different criminogenic needs – appears to hold significant promise. The successful results of the Te Piriti evaluation, which demonstrated that a programme that combines tikanga Māori processes with Western psychology, should offer strong encouragement for this area. The Te Piriti programme has demonstrated that a tikanga Māori component has an additive value over an exclusively cognitive behavioural approach to the treatment of child sex offending. Whether the same result could be obtained by stacking a tikanga Māori programme with a cognitive behaviour programme has yet to be answered.

Finally, many of the reviewed evaluations do not report how many offenders failed to complete programmes that they started, the reasons for their non-completion and what happened to them in terms of recidivism after dropping out (a notable exception being the Montgomery House paper by Berry, 1999). A number of studies, both internationally (Stark, 1992; Wormith & Olver, 2002) and in New Zealand (Department of Corrections, 2001b) indicate that treatment drop-outs have a higher recidivism rate than not only treatment completers and control groups, but in some cases also treatment refusal groups. The inclusion of data on treatment non-completers also potentially allows for the investigation of any possible selection bias (e.g., differences in risk, ethnicity, initial motivation) between completing and non-completing groups. Future evaluative studies should ensure that the issue of treatment non-completion is fully documented in the reporting of any results.