The influence of chronological period of data collection on differences in reported dietary intake among school-aged children surveyed in 39 countries

Objective: To examine whether dietary intake reported by school-aged children relates to the chronological period of data collection.

Design: Cross-sectional surveys in 39 countries in different monthly periods of the 2009-2010 school year.

Settings: Questionnaires were completed in schools anonymously.

Participants: Children from 39 countries, aged 11, 13, and 15 years (n = 209,320).

Main outcome measures: Daily consumption of fruits, vegetables, sweets, and soft drinks, derived from a food frequency questionnaire.

Analysis: Multivariate logistic regression (applied for 3 countries); 2-level random intercept logistic regression (applied for 36 countries).

Results: Monthly variations in food intake among students from Canada, England, and Norway, where data collection took place almost all months of the school year, revealed significantly lower daily consumption of food items in January to February. A 2-level random intercept logistic regression model for 36 countries, where questionnaires were administered in relatively shorter periods, indicated the lowest likelihood of fruit and soft drink consumption when the data were collected in March to April (β = -0.30; P = .006, compared with October to December) and in January to February (β = -0.65; P = .018, compared with May to June), respectively.

Conclusions and implications: The chronological period of data collection must be considered when comparing children's reported food consumption, but this effect is small relative to cross-national variations in dietary patterns.