Aggregate comparisons of self-reported versus nonself-reported drinking in a general population survey.

Insight is gained into the validity of self-reported drinking in the general population by comparing self-reports and nonself-reports on the aggregate level. Married and cohabiting respondents of a general population survey (N = 2,169) were asked about both their own and their spouses' drinking behavior. It was found that on the aggregate level, distribution of "moderate" drinking and usual frequency of drinking is similar between self- and nonself-reports. Self-reported "heavy" drinking, however, is lower than nonself-reported "heavy" drinking among women in general, older women, and women with a lower education. Among men in general and older men in particular, however, self-reported occasional "heavy" drinking was found to be higher. The similar distribution of "moderate" drinking and usual frequency of drinking between self- versus nonself-reports gives reassurance about the validity of self-reported drinking behavior. The discordance in self-reported versus nonself-reported "heavy" drinking, however, raises questions about the validity. Interpretation of the discordance is not conclusive: more research (experimental and qualitative) has to be done to disentangle this issue.