Evaluating Crime Prevention: The Potential of Using Individual-Level Matched Control Designs

The relative merits of quasi-experimental area-level place-based evaluations and randomized control trials are the subject of contemporary debate in the criminological literature. Pertinent issues in the application of these approaches to crime prevention include the experimental validity of the former and the real-world practicality of the latter. In this article, we propose a 'middle-ground' and discuss the use of matched control designs and survival analysis conducted at the individual level. This is illustrated using information relating to a target-hardening scheme implemented in the UK. Some advantages of this approach are that it enables the tracking of samples of action and control households over time, does not demand random allocation to groups from the outset and provides insights into the longevity of the impact of measures. Disadvantages lie in the extensive data preparation required and the dependency of the process on detailed records from agencies.