Realistic evaluation is the term given to a model of evaluation presented by Pawson & Tilley (1997). Essentially, they argue that it is not enough to know ‘What Works’, and that the question that policy makers and service delivery staff need to address is ‘What works, for whom, in what circumstances?’ This guidance supports this model, which emphasises the role of theory and the development of theory as a means of improving understanding. Pawson and Tilley argue that evaluators need to think through the whole process of why a programme might work, including the inputs, the outputs and the context in which the work takes place. They argue that it is the interrelationship between the factors that creates mechanisms for change and that without understanding the nature of the mechanisms, it is very difficult to understand why a programme is successful or not. This model is useful as it enables some understanding of the effectiveness of programmes with different types of offenders.