The apparent validity of diet questionnaires is influenced by number of diet-record days used for comparison

The effect of the number of diet records used as the reference data on the apparent validity of a diet history questionnaire was examined using data from 97 participants in the Women's Health Trial feasibility study. Pearson correlation coefficients were computed between the estimates of usual individual intake from the questionnaire and estimates from the mean of three 4-day records obtained over a 1-year period, the mean of the two most recent records, and the record obtained 1 year after the start of the study. The results illustrate that the apparent validity of a questionnaire increases when it is compared with a greater number of 4-day food records. The correlations increased for most nutrients when the baseline food record was included, despite the fact that it was completed at a time not targeted by the questionnaire. These findings suggest that greater effort should be given to obtaining more diet-record days when validating other dietary assessment instruments.