In this study we used a whole-room indirect calorimeter to evaluate the accuracy of self-reported food intake. Daily measured energy expenditure was compared with 2 weeks of self-reported food intake. Additionally, oxidation of each macronutrient was compared with its self-reported intake to assess the accuracy of self-reported dietary composition. Participants (23 through 60 years old) were eight dietitians, eight subjects who were trained in keeping dietary records, and eight subjects who were not trained. Physical activity in the calorimeter was matched to usual daily physical activity. Overall, measured energy expenditure was approximately 200 kcal/day higher than reported metabolizable energy intake. However, this was the result primarily of a few subjects whose self-reported food intake was considerably below measured energy expenditure. Subjects who were trained in record keeping did not differ from untrained subjects. Dietitians had the lowest difference between intake and expenditure; none of them had the large discrepancies between intake and expenditure seen in the other groups. In all groups, there was a much greater discrepancy between self-reported intake and oxidation of each macronutrient than between self-reported total energy intake and expenditure.