The household component of the Nationwide Food Consumption Survey (NFCS) collected data on the foods that a sample household used in a given seven-day period. The basic vehicle for collecting this information was a food list that was organized at three levels for most food items. First, respondents are asked if they used any foods from a broad grouping of foods, such as vegetables. We refer to this as the food group level. Respondents who report using the food group are asked if they used any foods from narrower, more specific categories within that group, such as corn or lettuce. We refer to this as the food category level. Finally, respondents who report using these categories are asked to report the specific foods they used, such as iceberg, Boston, or leaf lettuce. We refer to this as the food item level.

Our initial research was based on the premise that foods were being under-reported because the list was not organized in a way that respondents thought about foods. Our goal was to improve the accuracy and completeness of respondents’ reporting of foods, and an important way to achieve that goal was to make it easier to recall foods. Another advantage of simplifying the respondents’ task is that it would reduce respondent burden. Our focus was on the content and organization of the food list.