Response Effects in Consumption Surveys: An Application of the Beta-Binomial Model to Self-Reported Drinking Frequencies

Abstract

Prospective diaries, seven-day recall interviews, and a self-assessed summary measure are compared as methods of measuring drinking frequency. The distribution of the number of drinking days in diaries is modelled by the beta-binomial model and applied to a Dutch 1985 general population one-week diary and to a 1983 two-week diary. The model-based comparisons show that the deviances between the methods are most pronounced in respondents with intermediate drinking rates. Memory effects are suggested as an interpretation of the deviances.