ABSTRACT
There are various approaches, both qualitative and quantitative, to evaluating group programmes for offenders. In quantitative research the basic concerns lie in deciding upon the offender sample, choice of independent and dependent variables, length of follow-up post-evaluation, and research design. There have been various attempts to classify the reliability of qualitative research designs generally, alongside debates specifically as to the optimum method by which to evaluate group programmes. In examining the arguments for and against the numerous designs, the point is made that choice of design is critical with respect to the outcome of the evaluation. In particular, an Intention to Treat design may mask the effects of a successful programme, giving a null finding for the wrong reasons. It is suggested that applied research is best served by triangulation of the findings of research, which has employed a range of robust designs and associated methods of data analysis.
