Evaluation strategies for SNAP/EBT programs at farmers' markets

Many farmers' markets are interested in improving access to fresh produce for SNAP (food stamp) participants by adding electronic benefits transfer (EBT) capabilities at the market. Evaluating the impact of new EBT capabilities on SNAP/EBT sales in a rigorous way is a critical but often overlooked component of EBT projects. In this study, we review the analytic strategies and results of an EBT pilot program at the Clark Park farmers' market in Philadelphia, a market that accounts for approximately one-quarter of all SNAP/EBT sales at farmers' markets in Pennsylvania. The pilot replaced a single market-operated wireless point-of-sale terminal with multiple vendor-operated terminals for nine months in 2008-2009. The study uses time-series regression models and controls for both seasonal effects and for the total amount of SNAP benefits issued in Philadelphia. Models also test for placebo effects (i.e., increased sales before the start of the pilot intervention) and for the persistence or "stickiness" of the intervention after the pilot period.