Data from state telephone surveys of self-reported seatbelt use, driving while intoxicated, and drinking five or more alcoholic drinks at one sitting were compared with objectively observed belt use in traffic and evidence of blood alcohol in fatally injured drivers. Self-reported belt use overstates actual use by more than 20 percentage points on average. Self-reported alcohol use is not predictive of the percentage of fatally injured drivers with evidence of blood alcohol among the states.