This article is part of Busara’s Tafakari 2026 yearbook. You can download the whole yearbook here.

During the 2022 FTX Crypto Cup, the chess world accused Hans Niemann of cheating. In one of the interviews after he won a great game, he was offered the mic, and he said a line that quietly lodged itself into my brain: “chess speaks for itself.” The point was simple. When confidence in fairness collapses, the best defense is not louder persuasion. It is the quality of the moves on the board.

That idea feels especially relevant right now. The development world is going through a rough patch. Funding is tight. Trust in evidence is strained. Many organizations are asking whether careful research still matters. At Busara, our answer is not rhetorical. Our answer is the work. We are doubling down on the fundamentals of sound behavioral science so that, when the mood shifts, the integrity of our results will already be standing tall.

Idea one: Innovation without cutting corners

Recruiting couples into a behavioral lab is rarely easy. We are essentially asking individuals to take part in a behavioral experiment (this recruitment can already be tough) and then do it with their partner (which might inadvertently seem like their partners might find out things about them that they did not want to share). This became even more complicated in one financial privacy study in Nairobi where part of what we were trying to understand had to do with how couples interact, meaning we had to ask real spouses to make decisions about whether to reveal or hide income from their partner. 

The challenges pushed us to think harder about what ethical and rigorous recruitment looks like when incentives exist, and life gets messy. Couples show up or disconnect from a study for many different reasons. We quickly learned that within-couple attrition (meaning one partner dropped out of the study) could compromise the value of our collected data if one partner failed to show up.

What to do when reality hits? Design for reality. We invited men first because that increased the chance that their wives would join. We allowed couples to arrive together or separately, respecting childcare rhythms. We offered a small bonus upon completion of both spouses’ data collection. There was a slightly higher bonus for showing up together at the lab, which nudged attendance without making the decision hostage to travel logistics.

Arrival solved the turnout problem but introduced a second one. Not every “spouse” was a spouse. In Kenya, marriage does not always come with paperwork, so some participants saw an opportunity. We responded with respect and a little curiosity. Screening questions covered years together, where they met, and how many children they had. The survey included additional checks like who wakes up first or matching mobile numbers. Eight fake couples tried their luck. We thanked them for their time and reimbursed their transport home.

Even once verified in their coupledom, privacy still needed protection. Couples sometimes sat near each other, which meant choices about income could become performative. We seated partners far apart. We used random lotteries to prevent anyone from easily inferring the other person’s confidential decisions. Only public decisions were disclosed by text. That quiet design thinking protected dignity and produced data we could trust.

Idea two: Slowing down to remove weak links

We believe in moving quickly. We also believe that some pauses are worth the cost. Our Enumerating Development study has a clear purpose. We want to understand whether flattening hierarchies between principal investigators and field officers can improve motivation and data quality. Enumerators often report feeling like robots.  They worry that poor performance is punished and good performance is invisible, and that their work is defined more by the number of surveys they complete in a day than by the quality of the data. Their working conditions are often poor, and they feel intellectually intimidated working closely with project principal investigators (PIs).

All of these behaviours are invisible in standard protocols, yet they shape truth on the ground. That is why we designed a two-country experiment involving hundreds (if not a thousand) field officers. Some will receive standard training, and others will participate in activities that strengthen communication, nominate champions, and build trust with PIs. The goal is to determine whether enumerator agency improves honesty in data collection.

We delayed our submission to the Institutional Review Board (IRB) to get this right (and in 2026, Busara will launch its own ethics review board, which Joel Mumo writes about in this Tafakari). The delay increased costs. It also clarified ambiguities, tested the ethics of deception in the design, and provided stronger checks for data tampering and gaming. The decision was difficult in a moment when every shilling counts. We chose rigor because good process today prevents fragile findings tomorrow. (IRB reference number: USIU-A/ISERC/US1170-2025)

Idea three: Fewer mistakes. More transparency. Closer partnerships.

Tight funding does not excuse shortcuts. It requires discipline. We will not reshape hypotheses to please a funder. We will not search for significant results like a fisherman tossing back the small fish. We will publish methods openly, show uncertainty honestly, and call out the limits of what we know. The loudest failure is the one discovered later.

At Busara’s annual research festival, Tara Mistari (which Hitha M. writes about in this Tafakari), we even held a session called ‘Tracing the Ripple’ in which teams designed terrible research and tracked the chaos it created. People who skipped pre-testing ended up with surveys that no one understood. People who eliminated enumerator checks produced fabricated data. The humor made the lesson sting a little more, but the lesson was clear: Small errors ripple outward.

We want funders directly inside these conversations with us. We want them to argue over trade-offs early rather than late. We want to prove that a careful process pays returns. Collaboration prevents waste. Transparency builds trust that endures dips in confidence that everyone is feeling.

The path ahead is not glamorous. It has a lot of clean code, grounded consent procedures, documented protocols, and clear communication. And yet, that’s our path because we believe that when politics and finance recover and reunite to help build a better world, people will remember who kept their heads. As Niemann said, the game speaks for itself. We plan to let our research do precisely that.

Gitanksh Sethi is an Associate in the BRACE (Behavioral Research and Academic Engagements) team whose passion for community and collaboration inspires us all.