This study attempts to overcome some of these methodological difficulties by measuring deterrence in a drunk driving enforcement program in a small Wisconsin city. It demonstrates a statistically significant relationship between a carefully designed enforcement program and driver response measured by accident rates. This finding is contrary to some recent studies challenging the efficacy of enforcement efforts aimed at traffic accident reductions. A general discussion of deterrence and enforcement amplifies the findings and conclusions. A word of caution is added regarding the findings. (Author abstract)