All U.S. states have laws designed to discourage people from drinking and driving, but enforcement varies across the states. Existing studies offer conflicting evidence on the effectiveness of these strategies in deterring drinking-and-driving behavior. Deterrence theories imply that the mere existence of such laws has little impact on criminal behavior, but the perception of enforcement and the probability of being detected have a deterrent effect. To test these hypotheses, we develop a measure of the propensity to drink and drive using item response theory and national survey data. Inferential models test the impact of perceptions of enforcement, actual enforcement levels, and deterrence laws on criminal propensity. Results indicate that the existence of statutes impacts only those least likely to drink and drive, while perceptions of the likelihood of arrest and individual agreement with the goals of drinking and driving laws significantly reduce the propensity for almost everyone. Actual enforcement rates display no behavioral effect.