December 6, 2004: Review of 'What Works'

We need to know what works to reduce re-offending in order to decide what interventions to deliver to offenders. However, simply knowing ‘what works’ only identifies what should be delivered rather than whether or not it will be effective. A new report, "'The impact of corrections on re-offending: a review of ‘what works’ " focuses on correctional services interventions with adult offenders, aims to update and review knowledge of ‘what works’ in corrections.

Without knowing how interventions work, NOMS cannot maximise efficiency or effectiveness, as good practice cannot be identified and weaknesses eliminated. This review of the evidence therefore aims not simply to describe ‘what works’ in corrections but also to provide direction to build on the evidence of what works for whom and why.

If we accept that offenders have a wide range of criminogenic needs, then successfully  tackling them ultimately depends on an effective assessment system to identify them and measure change in the degree to which they are present. There is evidence to support the use of offending behaviour programmes, though the majority of this comes from meta-analytic studies and primary studies of research done abroad. In Britain, the evidence is mixed and limited.

The methodological constraints of British evaluations to date have meant that often it is difficult to separate the effects of programmes from other important factors that influence offending behaviour, for example, offenders’ motivation. These evaluations have also highlighted the difficulty in implementing offending behaviour programmes on a large scale.

The emerging evidence on basic skills training in prison suggests that these courses can improve prisoners’ skills but the extent to which these can be improved sufficiently to have a positive impact on employment prospects by prison training alone is still in doubt.

The evaluations to date of drug treatment programmes suggest that these programmes can reduce reoffending. The research also suggests that combining drugs treatment with cognitive behavioural interventions, particularly for less educated offenders, can increase the impact of the programme. However, the evidence makes clear that the gains made in prison can be quickly lost if there is insufficient aftercare for prisoners once they are released.

Harper and Chitty point out that evaluations of correctional services interventions have often been based on sub-optimal research designs. To help to tackle the use of sub-optimal research design, the adoption of an integrated model of reconviction is proposed together with ways to improve the measure of the impact of the correctional services.

However, Harper and Chitty argue, it is also important to recognise the reality that underpins offending behaviour. Offenders often have multiple criminogenic needs. As such, the research effort needs to reflect the multiple and complex problems of offenders.

It is important to examine the breadth and range of interventions that offenders receive in the context of these multiple needs and not simply to examine each intervention or need in isolation.However, to assess the impact of those interventions on re-offending, there is also a need to develop randomised control trials in the correctional services. This will ensure, conclude the authors, that knowledge of ‘what works’ is improved and the existing equivocal evidence is replaced with greater certainty and greater confidence in NOMS’s delivery of successful interventions to reduce re-offending.

The full study can be downloaded here.