Measuring Food Security in the United States for More Than 25 years: History, Methods, Findings, and Opportunities

Abstract

This article sets the stage for the “25 Years of Food Security Measurement: Answered Questions and Further Research” conference, with support from the Economic Research Service of the US Department of Agriculture, by providing some history of federal food security measurement, summarizing notable findings, and reviewing selected special topics in analysis methods. The federal government uses food security surveys to monitor national progress toward reducing food insecurity and to evaluate federal nutrition assistance programs. For the monitoring purpose, there is a tension between focus (on a single authoritative measurement approach) and breadth (encompassing multiple tools or instruments suitable for diverse populations, contexts, and applications). For the program evaluation purpose, challenges include coordination with study designs capable of real causal estimation in the face of strong self-selection effects and tailored reference periods in survey questions that match the timing of program participation. Some analysis methods treat the food security survey items as distinct experiences of hardship, whereas others treat the food security survey items as windows on an underlying latent variable, a food insecurity score. The severity of food-related hardship may be assessed quantitatively by the number of distinct hardships reported, by the estimated value of a latent food insecurity score, or by the frequency of occurrence for sentinel hardships. Ongoing work investigates statistical approaches that are sufficiently simple for policy application and yet sufficiently flexible to accurately match the empirical survey evidence.