Comparing food literacy by grade, sex, and food education exposure: construct validation of the food literacy BITE scale

Background Food literacy provides a framework for understanding dietary behaviors in children and can inform priorities for and evaluate the effectiveness of food education interventions. The 4-dimension Food Literacy BITE scale is built on a child-centered food literacy framework and has been structurally validated in U.S. elementary schoolchildren. We conducted additional validity testing to assess the tool’s generalizability, ability to differentiate mean food literacy by grade, sex, and food education exposure, and the association of food literacy score with food group consumption frequencies. Materials and methods We administered the Food Literacy BITE scale and food frequency questionnaires to n = 690 4 th and 5 th -grade students from 10 elementary schools in Washington, DC, and Kentucky (52.5% 4 th grade, 55.8% 4+ years of food education, 52.8% girls). Stepwise series of multiple-group confirmatory factor analysis models assessed measurement invariance by grade, sex, and school-level food education history. We examined group differences in overall food literacy score and dimension scores using latent means and used multilevel modeling for the food education exposure analysis to account for clustering within schools. We used weighted regression with robust standard errors to investigate associations between food literacy scores and food group consumption frequencies, aggregated by school and grade ( n = 20). Results Item interpretation was consistent across groups, with full scalar invariance established by grade and partial scalar invariance by sex and food education history (4+years vs. new/none). Girls scored higher than boys in the “confidence in everyday food skills” dimension ( p = 0.03). Mean food literacy was higher in the group from schools offering 4+ years of experiential food education ( p = 0.03), driven by higher scores in the “valuing shared food experiences” ( p = 0.02) and “food systems and nutrition knowledge” ( p = 0.04) dimensions. In the multilevel models, only the difference in “valuing shared food experiences” remained significant. Conclusion Results support generalizable use of the Food Literacy BITE scale in the U.S. and areas with similar food practices and environments. The tool’s ability to differentiate between food education exposure history suggests it may be useful in identifying best practices for food education interventions targeting late childhood.