ABSTRACT
Treatments in prison and probation services have, in recent years, been guided by the ‘What Works?’ body of research. In the 1970s, there was a view that ‘nothing works’ in offender treatment (Martinson 1974), leading to a shift away from offender treatment. However, in the 1980s, researchers began to identify what did work in offender treatment. They were aided by the newly emerging statistical technique of meta-analysis, which permits the aggregation of data from individual studies, thus giving a more accurate representation of treatment outcomes. Over the past two decades, many meta-analyses of offender treatment outcome studies have been conducted. Along with information on factors that are predictive of recidivism (see Chapter 6), the accumulated information has led to the empirically based formulation of the Risk-Needs-Responsivity (RNR) model of offender rehabilitation (Andrews and Bonta 2003; Andrews et al. 2006).
