already done a considerable amount of micro-level research as they have worked to understand the benefits achieved by the Food Stamp Program. They should continue this research.

The panel concludes that relying on trends in prevalence estimates of food insecurity as a sole indicator of program result is inappropriate. To assess programmatic results, better understanding is needed of the transitions into and out of poverty made by low-income households and the kind of unexpected changes that frequently bring about alterations—for good or bad—in households participating in food assistance programs.