In a 1979 national survey of drinking practices, a sizable number of respondents gave the inconsistent information that they drank more alcoholic beverages in bars and at parties than they drank overall. The percentage of respondents making such errors was too large to dismiss, and instead, it raised the possibility that the mention of specific contexts in questionnaire items produces response distortions. This suggests that caution should be used in the interpretation of survey data about drinking in specified contexts.