RESUMEN
OBJECTIVE: To use cognitive interviewing and pilot testing to develop a survey instrument feasible for administering in the food pantry setting to assess daily intake frequency from several major food groups and dietary correlates (e.g. fruit and vegetable barriers) - the FRESH Foods Survey. DESIGN: New and existing survey items were adapted and refined following cognitive interviews. After piloting the survey with food pantry users in the USA, preliminary psychometric and construct validity analyses were performed. SETTING: Three US food banks and accompanying food pantries in Atlanta, GA, San Diego, CA, and Buffalo, NY. PARTICIPANTS: Food pantry clients (n 246), mostly female (68 %), mean age 54·5 (sd 14·7) years. RESULTS: Measures of dietary correlates performed well psychometrically: Cronbach's α range 0·71-0·90, slope (α) parameter range 1·26-6·36, and threshold parameters (ß) indicated variability in the 'difficulty' of the items. Additionally, all scales had only one eigenvalue above 1·0 (range 2·07-4·71), indicating unidimensionality. Average (median, Q1-Q3) daily intakes (times/d) across six dietary groups were: fruits and vegetables (2·87, 1·87-4·58); junk foods (1·16, 0·58-2·16); fast foods and similar entrées (1·45, 0·58-2·03); whole-grain foods (0·87, 0·58-1·71); sugar-sweetened beverages (0·58, 0·29-1·29); milk and milk alternatives (0·71, 0·29-1·29). Significant correlations between dietary groups and dietary correlates were largely in the directions expected based on the literature, giving initial indication of convergent and discriminant validity. CONCLUSIONS: The FRESH Foods Survey is efficient, tailored to food pantry populations, can be used to monitor dietary behaviours and may be useful to measure intervention impact.