We assessed the effects of socioeconomic status and amount of time mother and child were together during the day on the accuracy of the mother's 24-hour recall report of her 3- or 4-year-old child's dietary intake. Accuracy of the mother's report was compared for nutrients and foods with dietary intake recorded by an observer. Neither socioeconomic status nor amount of time mother and child were together were related to any of the nutrient indicators of accuracy. Although socioeconomic status and amount of time mother and child were together were not related to the indicators of food accuracy, only 64.8% agreement was obtained between mothers and observers on foods consumed. Mothers were more likely to underreport (17.7%) than overreport (10.4%) foods. The low mean difference scores for nutrient indicators supported the use of the 24-hour recall method when collecting group data. The substantial disagreement on foods, however, suggests that previous studies assessing agreement on nutrients have overestimated the accuracy of the 24-hour recall method.