Alex Coutts is a CTED graduate student researching economics in the developing world. He spent this past summer interning in Uganda working with on an RCT. Here are his experiences while working out in the field:
This previous summer I had the opportunity to help manage a team of interviewers in Uganda, while they conducted surveys for Innovations for Poverty Action (IPA) on a Randomized Control Trial (RCT) involving savings groups.
CARE has been introducing a particular variant of savings group – called a Village Savings and Loan Association (VSLA) to various countries in Africa, since 1991. In a VSLA,members make regular deposits, and anyone in the group is free to take out a loan, to be repaid with interest. This interest gets added to deposits, effectively making the group a “community bank”. CARE has been scaling up their VSLA programs all over Africa for years, however the IPA study is the first rigorous impact evaluation.
400 villages were randomly divided into treatment and control group and baseline surveys were done in all villages. I arrived about two years later – just in time to start a week-long training session with our (mostly university educated) local interviewers. The entire endline survey was done on netbooks, which makes final data entry much easier – but brings its own host of problems. (e.g. realizing on the last day of training that every netbook carried a virus which purported to delete all the data collected)
Once training finished and the netbooks were virus-free, we drove out West where we’d be spending the next 6 weeks moving from town to town every 3-4 days. Accommodation was pretty cheap by US standard ($2-5 per night); however some standard “western comforts” were often missing (hot water/electricity). To deal with unreliable electricity we brought our own generators to charge the netbooks. Part of my job description became swapping netbooks during the night (beside my bed) to ensure they were all charged by morning. The task became harder as the project went on as the netbooks power connectors started breaking (60% were broken by the end) – so 20netbook batteries had to be charged by 8 netbooks (that’s 3 rounds of swapping out batteries during the night).
To give an idea of the scenery – green/gorgeous rolling hills with Matooke (plantains) growing EVERYWHERE. It’s the staple food of the west, and people eat A LOT of it (our staff routinely ate 8-12 each day). It can also be distilled for makingliquor (which also may have contributed to their often being a token inebriated male in one of every few villages). I accidently made the mistake of letting it slip how beautiful the scenery was in a certain remote village – whereby I was corrected by the village chairperson who agreed, but countered with something along the lines of, “Try to grow crops on these hills, and they’re no longer beautiful”.
All in all, doing the actual surveys was the easiest part. It was logistics that made this a tough, demanding project. We worked 6 days a week, left at around 7 am in the morning, and returned between 7-9 pm at night. Getting to villages only 30-40kms away could take up to 3 hours due to poor roads/or simply no roads wide enough for a vehicle. We also used a local Matatu (taxi-van) rather than the preferred Muzungu mode of transport (White Land Cruiser). These things were incredible, and could seat easily 15+ people.
With data collection done, the cleaning began – but the results are still preliminary. As a sneak preview let’s say that there are some modest effects – people seem to be shifting their reactions to negative shocks towards informal lending arrangements and transfers, rather than selling assets/livestock or reducing food consumption. It’s only a first glance at the data, but I’ll follow-up once the actual analysis has been completed.These and other results will help CARE (and other NGOs) learn which of their interventions have the most impact, so they can make the best use of their resources.