Religiosity and the validity of self-reported smoking: The Third National Health and Nutrition Examination Survey

Based on theories of social desirability bias, it was hypothesized that smoking, which may be socially undesirable in some religious circles, might therefore be under-reported by more religious persons. To test this hypothesis, data were examined from a cross-sectional survey of Americans, the Third National Health and Nutrition Examination Survey (NHANES III). Participants were American women and men aged 20 years and over. Measurements included self-reported frequency of attendance at religious services and of cigarette smoking, and measured serum cotinine, a metabolite of nicotine. Among smokers aged 20 and over, under-reporting of smoking did not vary appreciably with frequency of attendance. An exception to this was non-African American men aged 20-59 and African American men aged 60 and over who showed greater reporting bias among infrequent attenders, even after controlling for socio-demographic variables. No evidence for greater underreporting of smoking among more religious persons was found.