A comparison of a self-administered dietary questionnaire and a short interview

Abstract

The reliability of a quick dietary questionnaire developed for epidemiologic purposes was evaluated. Twenty-four hour dietary recall data by self-administered questionnaire were compared with 24-hour requency of consumption of various food items were obtained for 75 subjects. For many food items there was no systematic error between questionnaire and interview data, as the mean frequencies were nearly identical. However, some typical snack foods had higher frequencies in the interview than in the questionnaire. The chance-corrected agreement between the questionnaire and interview data was high for everyday food items such as potatoes and wholemeal bread. The level of agreement was not acceptable for some items including snack food items such as coffee, cakes and buns. However, the disagreement was frequently within one frequency unit of the pair value. It was concluded that the short questionnaire method for 24-hour recall seemed suitable in epidemiologic studies where the diet of large groups is under study. However, if snacks are considered important, it is recommended that the questionnaire be checked in the presence of the subjects when they hand it over.