You Say Potato, I Say Vegetable; You Say Tomato, I Say Fruit: Cognitive Validity of Food Group-Based Dietary Recall Questions.

Background: There is a need for valid, standardized approach for list-based questionnaires to measure food group consumption for indicators of diet quality including dietary diversity. Objectives: A common method for collecting dietary diversity data consists of open-ended food group questions, e.g. "Yesterday, did you eat any vegetables, such as cucumber, cabbage, or celery?" We sought to examine the cognitive validity of open-ended questions that require respondents to categorize foods and closed-ended alternatives using sentinel foods. Methods: Pretesting and 83 cognitive interviews were conducted in 5 languages in São Paulo and New York City in 2018. In structured interviews, respondents were asked to describe their thought processes in answering each question. Their feedback and responses to closed-ended and open-ended food group questions were compared. The Gallup World Poll then piloted 2 versions of the questionnaire in a nationally representative sample of 1000 in Brazil in 2018. Results: Respondents in all settings miscategorized foods when asked open-ended food group questions (0%-82%, depending on the food group), respondents varied in their ability to think of other foods that belonged to specified food groups (35%-50% could think of any items), and open-ended questions presented an additional cognitive burden. There were no significant differences between the results from closed-ended and open-ended questions in the national pilot test. In the context of a multitopic survey, the finalized questionnaire took 3-5 min to answer, had no additional training requirements, and enumerators reported similar ease in administration as modules on other topics. Conclusions: For data collection on food group consumption, open-ended questions requiring respondents to categorize foods present cognitive validity problems. Closed-ended questions using sentinel foods reduce or eliminate ambiguity, presenting lower cognitive burden and greater comprehension. Based on these results, the closed-ended method has been adopted in international survey platforms for measuring dietary diversity and other aspects of diet quality.