This analysis examines whether roadside surveys that measure changes in the prevalence of driving after drinking can be used to evaluate the impact of interventions designed to reduce alcohol-related traffic crashes. Using data collected at a roadside survey over a 4-year period in two California communities (N = 33,614), this analysis examines the relationship of BAC and drinks since 5 P.M., aggregated on a monthly basis, with nighttime injury crash data from the California Highway Patrol aggregated on a monthly basis. A regression analysis was used in which daytime injury and drinking style were covariates. The data demonstrate that temporal changes in drinking and driving collected at a roadside survey reflect temporal changes in nighttime injury rates. Hence, this study substantiates the use of roadside surveys as an appropriate methodology to evaluate drinking and driving interventions.