Evidence-Based Programs: Results from Research in the United Kingdom
Evans, Donald G., Corrections Today
Regular readers of this column will recall that I have been interested in the efforts in England and Wales to reduce re-offending through the introduction of evidence-based programs. The late Sir Graham Smith, inspector of the Probation Services in England and Wales, concerned about the future of the probation services, commissioned a report to examine which programs and practices were working (reduction of offender re-offending) and which were not. In the foreword of that 1998 report, Strategies for Effective Offender Supervision, Smith said this was a very important period in probation because it provides "an opportunity to renew and revitalize community penalties ... The rewards will be immense in terms of increased confidence and public belief." When, as a result of this research, the probation service introduced programs based on the principles advocated by the "what-works" research, they included a research/evaluation program to test their own program initiatives. At the International Corrections and Prison Association conference, held in Edinburgh, Scotland, in October 2005, Peter Raynor, Ph.D., from the University of Wales, Swansea, presented a very interesting paper on the what-works experiment in England and Wales. His focus was on both implementation and evaluation plus the lessons to be learned from the effort to reform probation based on an evidence-based approach.
Raynor noted that this was the world's largest experiment in evidence-based correctional practice, in that it encompassed both the probation services and the prison system in England and Wales. The experiment has now lasted approximately 10 years, with large-scale targets for completion of offending-behavior programs. For example, in 2005, the targets were 15,000 offenders in the probation service and 7,000 offenders in prisons. They were to be enrolled in programs that focused on offending behavior and had demonstrated effectiveness in reducing re-offending. Raynor said that the what-works strategy for probation stressed two main areas: the integration of the separate areas of probation service into one national service and the development and implementation of pathfinder projects (programs geared to reducing re-offending with an evaluation component attached).
Raynor then discussed the first wave of evaluation results. In three evaluations of offending behavior programs implemented in a prison setting, the results were mixed. In the first evaluation (2002), the program showed a positive effect for offenders assessed as medium risk. The second study (2003) showed no significant effect on the offenders' behavior. In the third study (2003), positive effects for those who completed the program were demonstrated. However, all the studies demonstrated problems of matching comparison groups (e.g., ability to match only on static risk factors and not on dynamic risk factors).
The results in probation demonstrated the following:
* Offenders completing programs reconvict less than comparison group offenders.
* Offenders who fail to complete programs reconvict more often.
However, the studies also indicated very low completion rates, only 21 percent to 38 percent in the major studies, thus making evaluation difficult because of the lack of distinction between program effects and selection effects. There is an indication, he noted, that there is an improvement in completion rates, but with only a minority of offenders entering the programs completing.
Raynor emphasized that, although the results of these first studies are far from conclusive, some results are generally positive as an introduction of the what-works strategy. These results highlight that probation services are now committed to reducing re-offending and that the majority of probation staff now understand the principles of effective probation practice. He also noted that the results, although mixed, provide the probation service with significant lessons for theory, research and implementation. …