ABSTRACT
Being able to measure change is one thing, being able to account for it is quite another. Three bodies of literature informed the data analyses and, in particular, helped to make sense of change: ‘realistic evaluation’, as most notably outlined in Pawson and Tilley (1997); developmental and interactionist approaches to criminal careers (see, for example, Sampson and Laub 1993; Thornberry 1997; Ulmer and Spencer 1999); and David Byrne’s work on complexity (1998). Pawson and Tilley placed social contexts centre-stage in their manifesto of how evaluation research should be improved and, as such, their work resonates with the intention of exploring in greater detail the role of social and personal circumstances in fostering or preventing desistance. The interest in developmental processes has been one of the most exciting growth areas in criminology in recent years, and has been important in shaping both conceptions and knowledge concerning criminal careers (see, for example, Moffitt 1997). Ulmer and Spencer’s (1990: 107) critique of ‘criminal careers’ from an interactionist perspective formalised a set of assumptions which were useful when exploring probationers’ responses to probation supervision. Finally, Byrne’s (1998) work has introduced to social research an agenda and a set of ideas aimed at exploring the relationships between three or more variables, especially when these relationships are non-linear.
