The existing criminological literature contains few analyses of the deterrent impact of legislation, and few of these are methodologically sound. However, statistical methods that are uniquely suited for the analysis of deterrent impacts are now available in popular, user-friendly computer software programs. These methods give criminal justice researchers the opportunity to become more involved in the creation of criminal justice policy. In the Arizona example, monthly data for daytime and nighttime fatal and injury crashes from 1978 through 1983 were analyzed and modeled using a quasi-experimental design. The analysis showed that the extensive publicity campaigns resulted in a significant reduction in nighttime crashes involving injuries or fatalities. Tables, figures, and 19 references.