Randy Kramer, environmental economist and director of the Duke Global Health Initiative, teaches a valuable course at Duke University’s Nicholas School. The course is called Social Science Survey Methods, and it involved me, as Teaching Assistant, when I was in the first year of my PhD. This week he asked if I would return to the class in order to share some insights with his Masters students about primary data collection in the field.
Reminiscing about fieldwork is a delight, so I readily agreed. (It’s a delight in the way that one might feel satisfied about running a marathon, long after the pain is forgotten.) I settled on five key points, based on insights from fieldwork in Mali and Burkina Faso (2008), the Solomon Islands (2008), Tanzania (2010), Mexico (2015) and India (2015). These represent field experiences in lower-income, rural settings. Obviously these points don’t represent a comprehensive list, but they may complement the more general advice of published field guides.
Find yourself an artist
One of the trickiest elements of primary data collection is translating abstract concepts into practical notions in a way your participants can engage with and understand. Art and design can help do this, and chances are reasonable that you can hire someone locally who has the required talent. Have him or her render your survey or experiment subject matter in color. Draw diagrams. Use tangible examples in those diagrams. We are visual creatures. These are ways in which you can translate the abstract into something comprehensible and meaningful.
Note also that you and your assistants will improve your ability to explain abstract ideas as you practice your survey or experiment. I have been amazed at the complexity of the ideas that can be conveyed once you find the right wording, props, diagrams or examples. Often the greatest barrier is not literacy, education levels, or complexity, it is the means by which the experimenters express themselves.
Art and design help at the other end also , when presenting results, and convincing people of why your topic matters in the first place.
Pilot (and be flexible!)
Mental-models of a situation, no matter how carefully devised, can fall apart when confronted with evidence. In Tanzania, I was certain that farmers would not object to a (hypothetical) ‘bonus’ payment being made to the village development fund, should they (hypothetically) conserve their forests. After all, who eschews more money? Interviews following pilot rounds, however, taught me that the village development funds were considered corrupt, leading people to express firm opinions against the funds receiving that money. Much better to direct the (hypothetical) payment to a (hypothetical) NGO. Problem solved.
Randomization is hard
I’ve learned the hard way that insisting on a random sample can offend or annoy people. ‘Why can’t my friend Juan play? He is helping you, no? Why are you rejecting our willingness to assist?’ Here are two ways to get around this. The first is to do a lot of preparation. Spend time in the communities, get to know key people really well, create your own comprehensive community household or population registry, select from it, and communicate, over a period of weeks (time spent building the registry) why you’re being selective. Use dice to indicate true arbitrariness. Or, use a ‘every-third-house’ type strategy, again, explaining to everyone why you’re being picky. This takes time! Alternatively, and more successfully for me, invite everyone in a community to participate. If you’re just surveying heads of households, this is feasible in small communities. You can drop observations during analysis if required.
Where possible, ride a bicycle
I’ve had the experience of showing up in rural communities by different means of transportation: By canoe, in the Mali inner Niger delta; by 4WD in Tanzania, Mexico and India; and on motorbike, foot and bicycle, in Tanzania. In the latter country, I borrowed an old beach cruiser bicycle a few times. It was useless for going up hills, and was too small for my long, lanky legs. But the locals thought the sight of me on this ridiculous bike was hilarious, and the kids loved it. Right away it gave us all something to laugh about (me), making for a much easier, warmer, and more rewarding community visit. Contrast that to stepping out of a 4WD, which to me, feels like getting out of a spaceship. In a place where 4WDs are unusual, it puts distance between you and the community at the crucial moment of first impression.
Of course, logistics, budgets and safety considerations means that there will be limitations on how you travel. But if you get a choice, keep it down-to-earth.
Try to give back
People like to hear about their community. They like to hear about an outsider’s perspective on their community. If you can, organize lunch and hold an open meeting. Tell people what you found (in simple terms). You may not be able to share sensitive or detailed information, but even just some summary statistics or personal observations help show respect for your participants’ intelligence and curiosity.