Assessing Potential Technical Enhancements to the U.S. Household Food Security Measures

The statistical measures used by the U.S. Department of Agriculture since 1995 to monitor the food security of the Nation’s households—the extent to which they can consistently acquire adequate food for active healthy living—are based on a single-parameter logistic latent-trait measurement model (the Rasch model). A panel convened, at USDA’s request, by the Committee on National Statistics (CNSTAT) of the National Academies in 2003-06 recommended that USDA explore five potential technical enhancements to that model. USDA has adopted one CNSTAT panel recommendation, which corrects the methods used to model the frequency-of-occurrence followup questions in the food security scale. This study examines the implications of that change and assesses the other four potential enhancements and the extent to which they would affect USDA’s published food security statistics. The study findings suggest that introducing the more complex statistical models would improve measurement of food security little, if at all, while making results and methods more difficult to explain to policy officials and the public