Design, Sites, and Data Sources
In 2015, the U.S. Department of Agriculture’s Food and Nutrition Service competitively selected and funded 10 pilot projects; the projects operate in the following states: California, Delaware, Georgia, Illinois, Kansas, Kentucky, Mississippi, Vermont, Virginia, and Washington. MDRC is leading the RCTs in 4 of the 10 states — Georgia, Illinois, Kansas, and Virginia. As part of the evaluation, all pilot projects are receiving technical assistance with implementing the new E&T services from the evaluation team.
The evaluation team is rigorously measuring the effects of the new pilot services through a randomized controlled trial in each of the 10 states. In this design, SNAP applicants and recipients who are eligible for the new SNAP E&T pilot services will be randomly assigned to one of two or more research groups: a group eligible for the usual SNAP employment and training services; a group eligible for the new SNAP E&T pilot services; or, possibly, a group eligible for a somewhat different version of the new SNAP E&T pilot services. By comparing the employment, earnings, SNAP receipt, and other outcomes of the groups over time, the evaluation team will measure the extent to which the new SNAP E&T pilot services help people more than the usual SNAP E&T services, and determine whom the new services help most.
The evaluation will analyze a variety of data: baseline data collected at the time of random assignment, program activity data from each site, and administrative data (for instance, SNAP receipt data, welfare receipt records, and unemployment insurance wage records data). In addition, two follow-up surveys will be conducted for a subsample of study participants (one at
12 months after their study entry and second one at 36 months, with a possibility of a third follow-up survey at 60 months).
In addition, detailed information will be collected on the implementation of the pilot services through three rounds of site visit interviews with program staff and through participant focus groups. Collected cost data will include program financial data and staff time-use surveys.