Originally published on the BIT website on 26 February 2026
Many of today’s most pressing social and economic development challenges are rooted in human behavior. From curbing the overuse of antibiotics and ensuring school attendance to promoting sustained saving habits in households.
Even structural and technological solutions rely on behavioral assumptions. For example, a school latrine programme in India found that building toilets increased enrolment, with much larger gains for girls when separate, sex-specific facilities were provided.
Often large-scale projects like vaccination campaigns require a substantial focus on human behavior to reach the last mile of delivery and ensure no one is left behind. For example, BIT designed and deployed a behaviorally-informed chatbot, which supported thousands of people to navigate the health system, answer queries and ultimately obtain their COVID vaccination in Argentina.
Behavioral science is rooted in understanding how humans behave in the real world, to design and test interventions that can be replicated and scaled for social impact. Yet, we know that without a solid grasp of norms, values and mental models across contexts these efforts could fall short, with common pitfalls like low programme uptake and distrust turning shiny silver bullets into flopped pilots and spoilt funding.
This matters because governments around the world spend trillions each year on public services and development programmes, but spend very little on finding out what works. In our 2024 global evidence report, we estimated an annual $85-$115 billion gap in social science research outside health and defence across four high-income countries.
At BIT, we have delivered 2,000+ projects in over 100 countries, and behavioral science has become a familiar part of public policy. Yet the evidence base remains fragmented, and we miss opportunities to accumulate knowledge across organisations and settings. To address this gap between ambition and evidence, together with the Behavioral Science Group and the World Bank, we have co-founded the Global Behavioral Evidence Network for Development (GBEND).
Behavioral science to advance development goals globally
Behavioral science has proven time and again that it can support development goals globally. However, more can be done to ensure that our understanding of human psychology is truly rooted in global data, and that we have frameworks to understand how biases and interventions manifest in different contexts. In fact, many widely used frameworks and interventions were developed and tested in “WEIRD” contexts – that is, Western, Educated, Industrialised, Rich and Democratic societies.
While it is often assumed that many core concepts will generalise, this underestimates how strongly behavior is shaped by social, institutional, and cultural contexts. Social norms, for example, operate very differently in individualist and collectivist societies. Trust in government, state capacity, and enforcement also vary widely. These factors, and more, shape how behavioral interventions work. We need more evidence not only on the traversability of existing behavioral concepts but also on new concepts emerging from non-WEIRD settings.
A decade of behavioral science in development
BIT has spent over a decade applying behavioral science in low- and middle-income countries around the world (see the full breadth of our work here). We’ve worked with governments, multilaterals, delivery partners and communities across health, education, financial inclusion, gender, governance, and climate.
A consistent lesson from this work is that context affects more than the effect size of interventions. It can also change the effectiveness of the underlying mechanisms themselves. For example, interventions that rely on individual autonomy may behave very differently in settings where decisions are collective.
We think there are three foundational areas of behavioural science that remain underexplored:
- How well do common behavioral concepts travel across contexts?
- What new methods and tools are needed for behavioral science in development settings?
- What institutional arrangements and funding models best support behavioral science for development?
What GBEND is trying to do differently
GBEND was created to address these gaps. It brings together governments, multilateral institutions, funders, academic researchers and applied behavioral teams to focus on shared problems rather than isolated studies. It was founded in May 2025 to generate and share behavioral evidence that serves the global majority to advance shared development goals across contexts.
The network is jointly governed by the Behavioral Insights Team, the World Bank Mind, Behavior and Development Unit (eMBeD), and the Behavioral Science Group. This structure reflects a deliberate choice: to connect applied policy work, global development expertise and regional leadership.
GBEND’s focus is on building the conditions for better evidence: this includes shared research agendas, collaborative trials, and mechanisms for learning across organisations and regions. Initial areas of work include antimicrobial resistance and AI-enabled research for low-resource settings.
Looking ahead
GBEND’s first major convening took place on Tuesday 3 February 2026, alongside the World Government Summit. We heard from partners across governments, academia and the development sector, and set out a shared agenda for the year ahead.
For BIT, joining GBEND is part of a longer-term commitment: to help build behavioral science that genuinely serves the global majority. That requires stronger foundations, deeper collaboration, and a willingness to question assumptions about what works, where, and why.
Authors
- Deelan Maru
- Chiara Cappellini
- Mónica Wills Silva
- Rachel Coyle MBE
- Dr Serene Koh
