Gold standard or fool's gold? The pursuit of certainty in experimental criminology

This article assesses some of the claims made for experimental research in the field of rehabilitation of offenders. It suggests that both policy officials and evaluators have tended to over-invest financially and intellectually in a technocratic model of reducing reoffending that emphasizes programmes for offenders, and to under-invest in models that see the process as a complex 'people changing' skill. It argues that the complexity of this process renders it hard to evaluate using experimental methods of evaluation such as randomized controlled trials (RCTs). RCTs provide strong internal validity, but in complex settings offer weak external validity, making it hard to generalize from the experimental setting to other settings. The article suggests that the proper role for evaluative research in this field should be seen as building and testing middle-level theories about how best to change offenders' behaviour.