Drinking and driving is perhaps the most serious problem associated with heavy drinking among college students in the United States. In this study, the authors examined drinking locations prior to impaired driving in a college student sample. They administered the Impaired Driving Assessment to 91 college students identified as high risk for drinking and driving. Participants reported an average of 7.98 (SD = 7.67) impaired driving episodes during the past 5 weeks. Using a random effects model, the authors found that location was a significant predictor of blood alcohol content (BAC), F(6,89.6) = 3.62, p = 0.0029. After drinking alcohol at a party, students' average estimated BAC prior to driving (geometric M = 0.089) was significantly greater than all other drinking locations (geometric M = 0.033). The findings of this study provide insight into drinking locations prior to impaired driving and can be used by college health practitioners to develop appropriate interventions to reduce the magnitude of this problem. (Contains 1 note.)