Can the American Community Survey provide new insight into household food security? An illustration of cross-survey multiple imputation

Abstract

National food security data have been vital in raising public awareness and motivating policy. This paper uses cross-survey multiple imputation as a flexible way to assess within-state food security patterns that cannot be measured well with the Current Population Survey. Using a CPS-based model, we impute food security status to households in the much larger American Community Survey. We illustrate the value of this approach by showing how grouping households by demographic or geographic attributes can provide insight into disparities within states by race and ethnicity, as well as variation in food security across substate geographies.