Like a host of other researchers, I personally disagree with Henrich’s argument about the process and culture of WEIRDness. I find it reductionist and even patronizing about different groups of people around the world. In this blog we will focus on the more common descriptive use of the term, as “a narrow subsegment of the world population.”

This narrow subsegment dominates the field of published research. The research that the development community cites and applies comes especially from the US, UK, Germany and a few other countries. It’s by Western, usually white, researchers at leading universities within these countries about well educated, well-off people within them. Those facts are largely indisputable, and that’s the definition of WEIRDness we’ll focus on here.

And it’s not just study samples either. We can break behavioral research into four parts, each of which draws from a narrow (WEIRD) subsegment of the population.

• Agenda setters: The people that prioritize research and their cultural milieu 
• Researchers: The people who formulate and execute the research 
• Location: The place the research is conducted 
• Participants: The people who are researched in that location 

The field is making progress in broadening “the participants” and “the location” of behavioral research, but little progress in the other areas. Less than 1% of authors and research participants in top psychology journals are from Africa despite the continent making up 17% of the world’s population. Even in areas where the research is becoming less centralized to the US it is largely through the inclusion of WEIRD Europe. This problem is faced by social science at large and the development community more broadly and has not significantly improved over the last three decades.

WEIRDness directly affects our work because it shapes whether or not we can effectively drive beneficial behavior change in international development. Blindly taking behavioral insights and techniques developed from well-off college students in the US can mean that we do ineffective, or simply foolish, things. 

For example, here’s a story often told at Busara about our first experiment in the field. While trying to measure the effect of stress on decision making we turned to a well-known, and often used, technique that was originally developed in Germany. The Trier Social Stress Test uses  public speaking to induce stress by  requiring participants to make an interview-style presentation, followed by a surprise mental arithmetic test, in front of an interview panel who do not provide feedback or encouragement. Similarly, we had our participants give a mock interview in front of a panel of experts dressed in white lab coats. In western settings, this method produces a reliable and large impact on measured stress: speeches are stressful, and white coats conjure an image of an authoritative scientist. 

So why weren’t our participants stressed at all?

After speaking with our participants and our Kenyan colleagues we realized that speeches were not a major stressor for many Kenyans as public speaking is more common and expected than in the west. And white coats? They evoked images of a very different group of people – butchers.  One of our participants very directly asked “why are we being interviewed for a job by butchers?” This humorous and embarrassing first study helped us recognize (again) why local knowledge and context must always be taken into account before applying accepted psychological tools.

When we don’t understand the local context, we can ignore the broader impacts of our work. We usually have a particular outcome we want to achieve, and focus our analysis on that chosen outcome. While looking for ways to encourage women who reported domestic violence in Brazil to follow up for support, we were thinking of leaning on the power of reminders to close the intent-action gap. One Brazilian gender-based violence specialist on the team pointed out that it was actually a very potentially harmful idea. People in dangerous domestic situations in Brazil won’t have a private WhatsApp number or place to view it without being seen (i.e. local context). If their partner sees that communication, it can trigger severe violence (i.e., substantive context). Blindly taking lessons from existing research on using reminders would have meant risking women’s lives.

By not including local participants and their aims and desires in a human-centered way, international development can quickly fall into coercion or condescension.  The most ‘effective’ program in the world will falter if people see it as an imposition or insult to their dignity

Context, the king with a capricious kingdom.

So context matters. But which context matters when? I don’t know of any established framework in Behavioral Science for “which details matter.” Across the literature, we can identify that culture, personal experience, and the local decision-making environment all affect choices and behavior. Some authors are exploring genetics as well, but we’ll stay away from that here. 

Culture may create early-socialization trait differences in cognitive mechanisms, mental models, or activate group-specific mechanisms or responses like social norms.  Systematically different personal experiences (with localized tendencies) may activate different cognitive mechanisms or responses. The local decision-making environment may create persistent artifacts that activate different cognitive mechanisms or responses among people in that context.  

Each of these tell us that context overall should affect impact. Unfortunately, our limited and WEIRD research also means that our field is not in a position to make an empirically rigorous statement about the impact of WEIRDness. We have some good initial research on how particular mechanisms – like temporal discounting – may vary within and across countries. This research helps us imagine how specific interventions to overcome cognitive and behavioral obstacles might have differential impact across contexts. However, the number of cross-context impact assessments, linked to cross-context differences in cognitive and behavioral mechanisms, is truly scarce. Our lack of empirical data is compounded by the broader challenges of replicating research in psychology and social science overall. In a sense, the best way to quantify the effect of WEIRDness is to try to change it: create behavioral research that isn’t WEIRD and then compare the types of research and their outcomes to our existing body of knowledge. 

Expanding where we can 

Even without quantified impact, we can try to expand our research beyond WEIRD behavioral science for simple reasons of equity.  If you work in the field, you probably believe that behavioral science is a useful and effective tool for doing good in the world. This very (presumed) effectiveness, however, has an unintended consequence on equity. The research that powers our work is created by WEIRD people who, generally speaking, already have many resources to support their research, such as a high-quality education, well-funded research institutions, and grant funding options. Similarly, this powerful tool is primarily directed towards the interests – however benign and well-intentioned – of agenda setters within Western governments and philanthropies. 

To tackle WEIRDness practically, let’s return to the four aspects we talked about in the beginning: agenda setters, researchers, locations, and research participants. The latter ones are the easiest to address.  We start by understanding our own role.  Often, we discuss our work as “applying” behavioral science to international development and being “practitioners” who are distinct from academics. In reality, we are forced to also be researchers, because we can’t trust existing research and apply it blindly. Specifically, we are researchers with a unique set of locations and research participants; if we were to share our results regularly and rigorously, we could help address WEIRDness.  

To make this a reality, we must insist on evaluation: qualitative or quantitative (or ideally, both!) We also can’t stop at our target outcomes (uptake, usage, program impact); we need to be watching and searching for our broader impact on the community and potential unintended consequences. This doesn’t mean that we insist on being the ones who conduct end-to-end research; rather, we must be clear with ourselves and our clients that someone must do an assessment. Then, we should publish our lessons and tools, through platforms like the Behavioral Evidence Hub and the Science of Behavior Change. Anything else is a mirage: our field simply isn’t advanced enough to apply behavioral science in international contexts without a thoughtful evaluation of impact. Through these evaluations we can learn how to de-WEIRD behavioral science itself.  

Second, we can change how we do our design work in the field. Participatory design is a great start, in that it brings research participants into the obstacle identification and intervention design process: it enables them, partially, to serve as researchers themselves. It is incomplete, though: participation is still constrained and guided by an externally set agenda, and often by WEIRD researchers.  Where possible, we should start with local priorities instead of embarking on a behavior change program that we, or our funder, has decided is important.  We usually do not know what behavioral questions local participants would prioritize and pursue themselves, if they had the funding and tools to do so. 

Third, we turn to funding: that is where we are the most constrained, but not without influence. Behavioral scientists in academic institutions can often set the agenda and approach funders. Applied behavioral scientists, however, are usually brought in to solve a behavioral problem that the funder has already identified. As a result our work reinforces the agenda-setting power of funders: philanthropies, governments, and multilateral institutions – primarily from the US and European countries. These agendas are not malicious or badly designed; they are simply from a relatively narrow segment of the world’s population.   We should seek to broaden, not replace, this agenda. 

Applied behavioral scientists, as recipients of external funding, should actively seek to give feedback and inform that agenda and advocate for a broader perspective, especially from the Global South. We should applaud the growth of behavioral science groups within funding institutions like the World Bank, UN and IDB, who can help support a broader, inclusive, agenda. We should also support funder efforts to localize international development: despite all of its imperfections (of which there are many), localization is a step in the right direction: with local funding, we can better understand local priorities and research questions. We can also participate in and learn from convenings of Global South researchers such as the Center for Effective Global Action’s Africa Evidence Summit.

Finally, we can look at our hiring and talent development practices: to change the researchers and future agenda setters of our field. This is an area that we at Busara have struggled with, as have most international groups in behavioral science. We are based in Kenya, but have a broad mix of people from around the world – including a leadership team that is disproportionately from well-off and well-educated populations in the US and Europe (this writer included). 

As someone who has hired many behavioral scientists over the years, I can say that WEIRDness is almost always an implicit or explicit criteria: through a particular educational background (until quite recently, there simply weren’t any non-WEIRD behavioral science programs) or job experience (again, until quite recently, there weren’t jobs at non-WEIRD behavioral science organizations). 

At Busara, we’ve tried to address this by focusing our hiring on skills-based interviews, and building our networks to attract a broader range of behavioral scientists. We’ve also sought to hire people without a behavioral science background and then train and promote within, to develop skilled non-WEIRD researchers that reflect the places in which we work. These efforts are imperfect and incomplete, but useful and we continue to learn every day. In the soon to be published book Behavioral Science for Development: Insights and Strategies for Global Impact, which includes a version of this blog post, my colleagues Mareike Schomerus, Stanley Ngugi and Deqa Aden have written a chapter on building a global team. The book is free to download  – and their chapter is great. 

Conclusion

Behavioral Science in international development seeks to improve the lives of people all around the world, and it has seen great successes. The field is also evolving, as it seeks to broaden its focus and become more representative of the Global South: to address WEIRDness in our research participants, locations, researchers and agenda setting. Early efforts have greatly improved the representativeness of the research: by studying the impact of behavioral science efforts on populations in the Global South, for example.

The field, however, is still lopsided, especially in terms of setting the agenda for behavioral science in development, and in the teams working in the field.  Our WEIRDness  may create practical challenges that limit the effectiveness of international development: from less effective interventions, to unseen unintended impacts, to inequity in the use of this powerful tool.

Once we recognize that the WEIRDness of our field requires us to be both practitioners and researchers, we can play a central role in fixing that problem.  We can honor the people and research in our field today, and expand both to be more effective.  Our projects can expand the base of research participants and locations, if we share our work. Our projects can support and demonstrate the promise of localization among agenda setters. And as applied researchers, our staff and our perspectives can help broaden the pool of researchers and agenda setters over time.   

Yes, our research and our field are both WEIRD – but it needn’t stay that way.