OBJECTIVES: to assess the reproducibility of a questionnaire to obtain self-recall of 16 foods eaten in five meals on a typical day of the week in elementary-school children. METHODS: the study subjects were 342 schoolchildren from the first to fourth grade of elementary schools (two public and three private) in the city of Florianópolis, State of Santa Catarina, Brazil. Reproducibility was determined by comparing responses obtained on two separate days, within a 3-day interval. Analyses were stratified according to school grade (first and second versus third and fourth) and type of school (public and private) to assess: the percentage agreement (%CC) and the adjusted kappa statistic. RESULTS: in public schools, %CC values ranged from 69% (bread) to 84.5% (fish), and kappa values from 0.38 (bread) to 0.69 (fish). In private schools, %CC values ranged from 73.9% (bread) to 91.6% (beans) and kappa values from 0.48 (bread) to 0.83 (beans). In analyses stratified according to grade, %CC and kappa values were better for third and fourth grades compared to first and second grades in both public and private schools. CONCLUSIONS: the reproducibility of the questionnaire was demonstrated for elementary schoolchildren in private and public schools, but the former presented better results.
Questionnaire; Food consumption; Schoolchildren; Reproducibility
