This study uses calendar year 2011 point-of-sale transaction data from a leading grocery retailer to examine the food choices of SNAP and non-SNAP households . On average, each month's transaction data contained over 1 billion records of food items bought by 26.5 million households in 127 million unique transactions.

This study examined the feasibility of creating a data collection system capable of directly and automatically providing USDA with item-level data on purchases made by SNAP households. Data would be captured at the point of sale from purchases made using EBT cards.

The second Access, Participation, Eligibility and Certification Study (APEC II) included a follow-on report that provided statistically-derived state-level estimates of school meals erroneous payments. However, while APEC II provided a rough indicator of relative risk for groups of states (e.g., higher than average, about average, lower than average), it was not a state-representative direct measure, and creating actual annual measures of such erroneous payments at the state level using APEC methodology is cost-prohibitive. This report explores alternative approaches to developing measurement-based state-specific estimates that are responsive to year-to-year changes in the actual underlying rate in each state. It also provides cost and burden estimates for the implementation of each of these methods.

This report assesses that pilot and includes a qualitative evaluation of the vendor and ITO experiences and an assessment of pilot costs.