HOW CHILDREN REMEMBER WHAT THEY HAVE EATEN

Objectives To determine whether students could verbalize, within 11/2 hours, how they remembered items eaten at the school lunch; to determine whether the categories of self-reported retrieval mechanisms were similar for two interview styles, integrated and nonintegrated; and to determine the effect of the two interview styles on the accuracy of reporting items eaten by comparing reports with direct observation. Design Two styles of dietary intake interviews were compared with observed intake in a school lunch setting. Setting Two elementary schools in Georgia. Subjects/samples Eighty-two of 106 fourth graders from four classes volunteered; 24 (six per class) were randomly selected and assigned to an interview style. Students interviewed using a nonintegrated style verbalized how they remembered after they had reported everything eaten. Students interviewed using an integrated style verbalized how they remembered at the same time they reported eating each item. Both interview styles included free report followed by prompted report. Main outcome measures Reported retrieval mechanisms were coded into 13 categories. Five measures of performance (specific match rate, general match rate, intrusion rate, omission rate, and overall match rate) were calculated by interview style for free report and prompted report separately. Statistical analyses performed We analyzed the effect of interview style on the number of students reaching 100% accuracy after prompting and on accuracy of reporting condiments using Fisher's exact test. Results Most students could articulate how they remembered items eaten. Reported retrieval mechanism categories were comparable for both interview styles. Visual imagery, usual practice, behavior chaining, and preference were the most commonly reported retrieval mechanisms. Accuracy of free reports did not differ by interview style; however, the nonintegrated interview style produced dietary self-reports with fewer condiment omissions during free report and higher accuracy after prompting. Applications Determining what retrieval mechanisms children commonly use for remembering items eaten may help researchers design cues to improve the accuracy of dietary self-reports. More accurate dietary self-reports could markedly affect the many types of research that use dietary assessment.