The project will estimate separate Engel functions, with food spending and food security outcomes a function of total resources for FSP participants and nonparticipants, using the Current Population Survey (CPS) Food Security Supplement for 2001-04. Food Stamp Program Quality Control files will be used to estimate food stamp benefits for each household. Outcomes will be allowed to vary based on specific elements of food security as well as the food security score, testing the validity of the Rasch model assumption that variation in food stamps and resources have a parallel effect on all elements of food security. Whether the effect of food stamps on food security is achieved through increased food spending will also be tested. An economic choice theory framework will be used to interpret results, with extramarginality assumed to be a matter of degree because households may have food expenses that cannot be paid with food stamps.