Around the world, governments are under pressure to improve learning outcomes. There is now an unprecedented wealth of research, yet this rarely reaches the people making decisions. The sector needs to shift from producing evidence for governments to generating it with them. True progress lies not only in discovering what works, but in building systems that learn, adapt, and implement at scale.
Governments around the world are under pressure to improve learning outcomes. Thankfully, the global evidence base on what works in education has never been richer. Yet in education ministries and district offices, decisions are still driven more by habit, beliefs, politics, or urgency than by data (OECD 2022, Dercon 2024). The challenge now is not a lack of research, but the difficulty of turning what we know into something governments can actually use.
The field of education has evolved into a complex web of researchers, implementing organisations, funders, and governments generating, sharing, and applying knowledge. This marketplace of ideas holds enormous promise: when evidence informs policy, resources can be used more efficiently, innovations can scale faster, and reforms can be more effective and better matched to the needs of schools and learners. But like any market, it is imperfect. Supply and demand rarely meet neatly. Researchers produce insights that do not always fit policymakers’ timelines or priorities, while governments struggle to find, interpret, or apply the right evidence when they need it most.
Bridging the divide between education research and policy
On the supply side, incentives for policy-relevant research are still weak. Academic recognition often depends on publication rather than utility, and studies can take years to reach the people making decisions. On the demand side, policymakers operate under intense time pressure, shifting priorities, and frequent staff turnover. They need fast, actionable insights, but most evidence arrives in technical reports or dense academic language. Political realities add further complexity: even strong evidence competes with electoral promises, budget cycles, and entrenched bureaucratic routines (Head 2016, Oliver et al. 2014).
Recognising these tensions, many actors in global education are rethinking what it means to connect evidence and action. Instead of assuming that research will naturally “flow” into policy, they are creating mechanisms where researchers, funders, and governments can work together from the start, co-defining questions, testing solutions in context, and adapting programs as they scale.
The Global Schools Forum (GSF), for example, has experimented with a network of Impact at Scale Labs that supports local education organisations to partner with governments. In these labs, promising ideas are tested in real-world conditions before being scaled. Organisations learn to collect the right data, interpret it with policymakers, and adjust their designs accordingly. A leadership training initiative for headteachers in Kenya, for instance, refined its approach after testing small behavioural nudges to strengthen the model’s applicability and utility to district education officers. Such examples show that scaling success often depends less on the model itself and more on how organisations learn and adapt through evidence.
At the same time, Innovations for Poverty Action (IPA) has worked with governments to bring evidence generation and use inside policymaking structures through Embedded Evidence Labs. For example, in Côte d’Ivoire, an evidence lab co-created with the Ministry of Education has embedded research and data practices directly into the ministry’s planning department. Ministry staff act as intermediaries partnering with researchers to jointly identify policy priorities, design studies and data products, and interpret results to inform the implementation of major reforms such as national pedagogy programmes. Because this work happens within the bureaucracy (and not around it), findings are more likely to be absorbed into day-to-day decisions and budgets (the case study below provides an overview of the Côte d’Ivoire evidence lab).
Across countries, Embedded Labs enhance both the design and implementation of policy, as well as the government systems that sustain them. Rwanda’s Ministry of Education Lab, for example, tested and scaled the STARS performance-based teacher scheme and developed a maternity leave module within the teacher management system to ensure teacher replacements. The Philippines Department of Education Lab applies predictive algorithms to identify at-risk students and prevent dropouts, while Côte d’Ivoire’s EdLab strengthened a national school dropout prevention programme by improving its Monitoring, Evaluation, and Learning (MEL) system. Together, these experiences highlight a subtle but important shift: from producing evidence for governments to generating it with them.
Case study: The Ministry of Education and Literacy (MENA) EdLab – Côte d’Ivoire
Launched in 2024 with seed funding from the Jacobs Foundation, the Côte d’Ivoire EdLab was co-created by the Ministry of National Education and Literacy (MENA) and Innovations for Poverty Action (IPA) to embed the use of evidence directly within government structures. Housed in the Ministry’s Directorate of Studies, Strategies, Planning, and Statistics (DESPS), the Lab works across multiple directorates within MENA, with support from the Inspector General and the Minister’s Cabinet, and engages external actors such as donors, universities, and other partners.
Through a participatory process, the Lab identifies the most pressing policy priorities via a Research and Learning Agenda (RLA). Guided by this agenda, it delivers targeted services according to need – for example, technical assistance for adaptive scaling of evidence informed policies and establishing MEL systems for large-scale policy implementation, as well as research studies addressing urgent challenges such as student and teacher dropout.
The Lab’s vision is long-term sustainability. To achieve this, IPA has emphasised a capacity strengthening and learning-by-doing approach, with ministry staff now co-leading activities such as designing policy-relevant research, developing MEL plans, and organising dissemination events to inform major reforms like Côte d’Ivoire’s national pedagogy program to improve learning outcomes. The EdLab’s inclusion in the country’s National Development Plan marks a major milestone toward institutionalisation. In parallel, a Jacobs Foundation-supported research funding mechanism is being established to finance studies aligned with the RLA, further strengthening the Lab’s role as a driver of evidence-informed policymaking.
By working from within the government, the EdLab helps ensure that education reforms are not only evidence-based but also responsive to the country’s real needs.
The realities of evidence-informed policymaking
From our collective experience, evidence-informed policymaking is rarely straightforward. Even when solid research is available, its uptake depends on context, incentives, and relationships. Policymakers weigh evidence alongside other pressures: the cost of implementation, political visibility, cultural alignment, or the feasibility of nationwide rollout. What looks cost-effective per student may become unviable at a national scale; what works in one region may fail in another due to differences in teacher incentives or language.
Timing also matters. Good evidence arriving too late (or without a trusted intermediary to interpret it and consistently advocate for its use), often loses influence. This is why approaches that combine technical rigour with proximity and flexibility tend to succeed. When researchers and policymakers work side by side, they can identify ‘policy windows’ early, adapt designs in real time, and build the trust needed for evidence to shape decisions.
These lessons cut across contexts. Whether it is an NGO testing a new learning model or a ministry refining a national strategy, the key is to create systems that learn continuously rather than episodically. Evidence use is not a single event; it is a habit formed over time.
Building a more intelligent marketplace
If the first wave of evidence-driven education focused on producing rigorous studies, the next must focus on making them usable. This means investing as much in translation, relationships, capacity building, and change processes within government as in research itself (Best and Holmes 2010). Funders are beginning to recognise this by supporting intermediaries that connect local evidence-based innovators with governments, or embedding teams within ministries to facilitate the design, adoption, and implementation of evidence-informed reforms.
In practice, this reflects the supply and demand sides of the market for education evidence converging. Governments articulate the questions they need answered. Researchers co-design methods that are credible yet feasible. Practitioners, intermediaries (Labs) and funders provide feedback on what can realistically be implemented or scaled, advocate for adoption, and support implementation.
Organisations like GSF and IPA, are helping shape this evolution from different angles. One works to expand the supply of credible, scalable solutions; the other strengthens the demand and institutional capacity to co-design, adapt and use them. Philanthropic partners such as the Jacobs Foundation play a catalytic role, supporting the connective tissue: people, partnerships, and platforms that make exchange possible.
An imperfect but living system
The global marketplace for education evidence remains imperfect. Research is unevenly distributed, incentives are misaligned, and politics will always shape how knowledge is used. But the fact that governments, funders, implementers and researchers are negotiating these tensions is itself progress.
The task now is to keep this process productive: to continue testing, adapting, and building the relationships that allow evidence to travel the last mile, from publication to policy to classroom practice. The future of education reform will hinge not only on discovering what works, but on ensuring proven solutions are implemented at scale and with high fidelity.
References
Best, A, and B Holmes (2010), “Systems thinking, knowledge and action: Towards better models and methods,” Evidence & Policy, 6: 145–159.
Bonargent, L (2024), “Political economy of education reform,” Unpublished manuscript.
Cairney, P, and K Oliver (2017), “Evidence-based policymaking: The politics of evidence,” Unpublished manuscript.
Damba, M, N Mtshali, and M Chimbari (2022), “Bridging research and policy in African education systems,” Unpublished manuscript.
Dercon, S (2024), “The political economy of economic policy advice,” Journal of African Economies, 33(Supplement_2): ii26–ii38.
Head, B (2016), “Toward more evidence-informed policymaking,” Unpublished manuscript.
Murunga, V, et al. (2020), “Evidence uptake in low- and middle-income countries,” Unpublished manuscript.
Nduku, A, et al. (2025), “Institutional incentives for evidence use in Africa,” Unpublished manuscript.
OECD (2022), “Who cares about using education research in policy and practice?” Unpublished manuscript.
Oliver, K, et al. (2014), “Barriers to evidence use in policymaking,” Unpublished manuscript.
Stewart, R, et al. (2022), “Knowledge translation in global education,” Unpublished manuscript.
Vogel, I, and M Punton (2018), “Learning about evidence in adaptive management,” Unpublished manuscript.