The accurate measurement of alcohol consumption in natural environments is a focal empirical issue in studies of the determinants of drinking. Four studies investigated agreement between subject and collateral questionnaire reports of drinking by adults over age 60 years (N = 83), and three of these studies also assessed agreement between daily subject and collateral prospective reports of subject drinking over 4 (Study 2) or 12 (Studies 3 and 4) week intervals. Subject and collateral questionnaire reports were significantly correlated. Excellent agreement was found for the group of subjects for self-monitoring reports of the number of drinking days and quantity of alcohol consumption per day, including reports of the temporal patterning of consumption, but variability in agreement existed across subject-collateral pairs. The discussion focuses on the conditions conductive to accurate measurement and on the importance of investigating predictor variables of individual differences in the accuracy of verbal reports of alcohol consumption.