This study tested the hypothesis that the intensity of behavioral impairment under alcohol is related to social drinkers' expectancies about impairment and their rates of rise in blood alcohol concentration (BAC). After subjects (n = 30) were trained on a psychomotor task, they rated the impairment they expected from alcohol and then performed the task under alcohol (0.56 g/kg) or a placebo. Alcohol impaired performance, compared with placebo. Drinkers' expectations about impairment and their rates of rise in BAC were independent, and each factor predicted a significant portion of the variance in alcohol impairment among drinkers. More intense impairment under alcohol was associated with expectations of greater impairment and with swifter rates of rise in BAC. BACs obtained by drinkers during task performance were not related to the intensity of impairment they displayed. The study shows that a pharmacological and nonpharmacological variable can each affect a drinker's behavioral impairment under alcohol, and this finding contributes to our understanding of conditions where BAC per se may be an unreliable indicator of impairment.