Interview format influences the accuracy of children's dietary recalls validated with observations

This report describes a comparison of the accuracy of children's dietary recalls obtained using either open or meal format interviews. Fourth-graders were randomly selected, observed eating school meals (breakfast, lunch), and interviewed that evening regarding that day's intake with children randomly assigned to open (n = 12) or meal (n = 11) format interviews. The weighted numbers of items observed eaten did not differ by format, but greater weighted numbers of items were reported eaten with meal format interviews than with open format interviews. Reporting performance was more accurate with open than with meal format interviews: Although the omission rates did not differ significantly between the formats, higher intrusion rates and total inaccuracy were found in meal format interviews than in open format interviews. Interview format influences children's dietary reporting accuracy; in particular, providing meal cues elevates false reports. These analyses exemplify the importance of treating omissions and intrusions separately when examining reporting performance. (C) 2003 Elsevier Inc. All rights reserved.