The Egyptian Social Fund for Development (SFD) and one of its microfinance institution counterparts identified 25 villages in which they were equally interested to expand microfinance operations. Using a pair-wise randomization process, researchers assigned these villages to treatment and control groups. First, these 25 villages were stratified into 10 sets of either two or three villages based on population size and region. Next, researchers assigned a random number to each village. The villages with the lowest number in each strata were assigned to the treatment group. This process created 10 treatment and 15 control villages with balanced populations and a balanced spread throughout Menya. This village identification and treatment assignment was intended to be replicated in 3 other regions. Ultimately, SFD and its microfinance institution counterparts decided not to expand lending due to political instability, economic uncertainty, and unexpected lack of capacity.
From the 25 villages identified by SFD and one of its microfinance institution counterpart, the team conducted a listing survey of at least 500 households per village. Villages had a mean of 111 business owners with 39 standard deviation. In villages with more than 100 business owners, 100 of the business owners were randomly assigned to participate in the household and enterprise surveys. In the 12 villages with fewer than 100 business owners, the team continued sampling until 100 business owners were reached. In total, researchers listed 13,413 households and completed 2,499 household and enterprise surveys for 2,525 business owners (26 households had two business owners).
