random sampling: it takes a village
For a pilot study on social networks, ethnic diversity, and collective action, I randomly sampled adults from four Malawian villages. Because participation in the study could generate cash earnings, the research team took precautions against creating jealousy or suspicion in the community — and I thought one good way to do that was to publicly draw the random sample. This is the story of how that sample was drawn.
It began with a village meeting. First, we were introduced by the headman and in all but one village (a predominantly Muslim village), there was an opening prayer. Then the Malawian research supervisor gave a 30-minute presentation that discussed the goals and process of random sampling and the purpose and nature of the study.
I was particularly impressed with his discussion of random sampling. He posed a question to the audience: if he wanted to know how the papayas in the village tasted, would he have to eat every papaya from every tree (pointing to the nearby papaya trees)? Some villagers laughed, many said “ayi” (no) aloud. He said, instead he would eat one or two from one tree, then take from another tree, but probably not take one from every tree in the village so that he could know more about the papayas in this village.*
After the presentation was over, villagers asked questions and we answered them. The meetings in each village were well attended, and in one of the villages, the meeting was rather celebratory in nature, with groups of women singing and dancing when we arrived. In all villages, women sat separately from men, but questions and participation in the draw came from both sides.
Once it seemed the village had a good grasp of our proposed work, we drew the sample. In an opaque bag, we had the name of household heads from the village representing each adult living in the village. For example, if Abiti Jamusi’s household had three adults, Abiti Jamusi’s name would be written on three pieces of paper in the bag.
Volunteer villagers would blindly select from the bag one piece of wadded-up paper, which would be unfolded by the research supervisor and read aloud (usually twice), so the person – or another member of the household – would make him/herself known. Everyone selected into the sample seemed happy to be chosen.
The villages averaged 200 adults in total, from which 55 were drawn into the sample. (Overall, then, we had 55*4=220 respondents in the study.)
Within households, a household representative would draw a numbered card for each of the adults sampled, and whatever number was drawn would determine which adult in the household would be sampled into the study. For example, at Abiti Jamusi’s household, each name was written on a numbered list and if Abiti Jamusi’s household had only one draw from the village sample, a research assistant would go to Abiti Jamusi’s home and ask that she or someone else draw one of three cards. Whatever number was drawn would determine which person on that household listing, whether Abiti Jamusi,** her husband, or her sister listed as 1, 2, and 3 respectively would be drawn. So, if #2 was drawn, Abiti Jamusi’s husband would be the sampled participant.
We cannot be certain, but we believe the transparent nature of the random selection and its conduct in a public space with the active participation of the community increased our ability to work freely in the villages without problems. In previous work with some of the same members of the research team, we encountered (not always friendly) questions from non-sampled community members about why they were not included in the study, and qualitative observations from the earlier study suggested some animosity between community members, especially when participants stood to gain something from participation. It might not have made much of an impact, but the time it took was short (half a day per village), and we would have had village meetings anyway to discuss the nature of the study.
I wasn’t the first to randomize in public when conducting research in Africa. I expect I won’t be the last.
* The analogy is his own making — he’s available for hire, by the way. See Hastings Honde in this video about a research group in Malawi.
** Abiti Jamusi is not a villager in the study areas. Rather, it is one of the monikers I take on when in Southern Malawi.