Assessing consumption of high-fat foods: the effect of grouping foods into single questions

Questionnaires on the frequency of consumption of foods are commonly used to measure dietary intake in epidemiologic research. To reduce the burden on respondents, questionnaires are often shortened by combining inquiries on similar foods into a single question. The effect of this practice on the reporting of dietary intake has never been investigated, however. To address this issue, we used two food frequency questionnaires in a telephone survey designed to rank adult residents of Alabama by their intake of dietary fat. One questionnaire included 29 questions about separate high-fat foods, whereas the other grouped these same foods into 14 questions. Compared with the 443 respondents interviewed using the 29-item separated-foods questionnaire, the 465 respondents responding to the 14-item grouped-foods questionnaire reported lower average intakes of the foods. In addition, a substantially higher percentage of respondents to the grouped-foods questionnaire reported never consuming the foods.