Ask a Behavioral Scientist: Piyush Tantia of ideas42 on the Importance of Context

This is part of our “Ask a Behavioral Scientist” series, where we give readers the opportunity to pose a question to leading behavioral scientists. The question below comes from a reader who works in the field of international development. Have a question? Ask it here.

Q: We are running the same development program in two different countries. It involves working with small farmers to improve the quality and quantity of their crops. But the results from Country A have far surpassed those of Country B. Under what seem to be similar conditions, Country A's farmers have improved their crop productivity much more than the farmers from Country B.

While not all of this may be attributable to behavioral factors, we would like to take a look at it through that lens. But we're not sure where to start and how to understand what behavioral mechanisms might be at work. Any suggestions for what questions to ask and how to go about identifying the "irrational" behaviors that might be derailing results in Country B?

As you say, there are probably many factors, not all easily visible, driving the differences between Country A and Country B, especially if you’re not comparing the results of randomized controlled trials (RCTs) in each of the two countries. Without RCTs, it is hard to know whether the development program, or related behaviors barriers, is driving any difference between the countries. It may well be that other factors are driving any differences.

That said, because your development program is having an effect and presumably requires farmers to change their behavior in some way, behavioral factors are probably present. To identify those behavioral factors, ideas42 would go through our diagnosis and design process over a few weeks. What follows is a very high level description of that approach. First, I would confirm that we have the problem right. In other words, is adoption of better farming practices in fact the bottleneck in Country B? If it is, we should find that interviewing and observing a sample of farmers in both countries reveals that those in Country B are not adopting new practices consistently. Chances are that you will see that farmers in Country B are only adopting a subset of the new practices, or are implementing them imperfectly. The more precisely we can identify the problem, the better.

Then we move on to diagnosing the problem. At ideas42, we have completed several projects in agriculture and there are many possible hypotheses. For example, we usually find that while farmers may “adopt” a new farming practice, they may ignore critical details like measuring accurately for spacing rows or burying fertilizer vs. leaving it on the surface. After generating some hypotheses based on our past experience and the psychology literature, we would conduct qualitative research to get some clues about which hypotheses may be right. Most importantly, we will want to investigate what in the context of Country B versus Country A may be triggering these behavioral issues.

At this stage, we generally ask “how” questions rather than “why” questions. For example, we would ask farmers how they went about planting and tending to their crops rather than asking why they did not follow the new practices. Then we would ask whether they considered any changes, how they typically learn of new practices or inputs, have they heard of any new ones recently, etc. This line of questioning will tell us whether farmers even intended to or remembered to try anything new in the first place. It will also give us insight into how they adopt new practices and which information sources they rely on. Doing this same exercise in both Country A and Country B may reveal differences that were previously hidden. Finding these precise differences in context will be critical because your program seems to be working in one country but not the other. Once we’ve investigated every hypothesis in this way, we would move on to designing possible solutions.

Typically, applying behavioral science insights would look something like this: hypothesize which psychologies may be getting in the way and then jump directly to known solutions to those barriers. For example, a behavioral scientist may think that the social norm around current practices in Country B is too strong, so the scientist designs a marketing campaign to communicate that the most successful farmers use different practices. This approach skips the critical “what” step: What contextual features are triggering the psychological barriers? Without that step, you cannot learn what is different in Country B vs. Country A. That “context reconnaissance” is the cornerstone of how ideas42 approaches behavioral design.

Have a behavioral science question on your mind? Submit a question to our network of behavioral scientists.

Disclosure: ideas42 is a founding partner of the Behavioral Scientist.