This study examined the extent to which food insecurity questions and the standard poverty measure were correlated with various dietary and physiologic outcomes. It used the National Health and Nutrition Examination Survey III, which between 1994-98 collected dietary recall information, surveyed respondents, and analyzed blood samples. Participants were nearly 34,000 people nationwide, with an oversampling of blacks, Mexican Americans, younger children, and older people to ensure adequate representation. Results indicated that the correlations varied tremendously by age. The food insecurity questions were correlated with dietary outcomes of older household members, but they were not consistently related to the diets of children. Even the food insecurity question that directly inquired about children skipping meals was not correlated with their dietary outcomes. In contrast, poverty predicted dietary outcomes among young children. Among adults, both poverty and food insecurity questions were good predictors of many dietary outcomes. The estimated effects of both poverty and food insecurity were strongest among the elderly. (Contains 11 references.) (SM)