Global Evidence Architecture: consolidating a fragmented landscape

Group Submission Type: Formal Panel Session

Proposal

The global evidence architecture is currently a fragmented patchwork of initiatives. Despite the large number of individual efforts, the education sector lacks a consolidated body of knowledge that national knowledge intermediaries like EdLabs and global funders such as multilateral banks and bilateral agencies need for evidence-based decision-making. This fragmentation hinders both global and local actors' ability to access, interpret, and apply research effectively.  

Based on an extensive sector consultation with key sector stakeholders- including two conveningsand commissioning of five thought leadership pieces – a persistent paradox emerged around the existing evidence landscape: while the volume of research continues to grow, a significant gap remains in consolidating, curating, and updating a knowledge base that national evidence translators—like EdLabs—can draw upon and contextualize for their countries in the form of synthesis (the systematic consolidation, analysis, and summarization of multiple research findings into coherent, actionable insights). Examples of synthesis include rapid reviews that quickly distil key findings for urgent decision-making, meta-analyses that statistically combine quantitative results, and systematic reviews that integrate diverse evidence streams to offer overarching insights.  

The process of producing an evidence synthesis, regardless of its form, is often multi-stage and resource intensive. It begins with the sourcing and searching of research, followed by abstract and title screening, data cleaning and deduplication, and then progresses to various layers of coding (including main, analytical, and quality appraisal). Only after these foundational steps can analysis and report writing begin. Timelines for production of synthesis can vary widely depending on the topic and type of synthesis (ie. rapid synthesis, meta-analyses, etc.). The process has been likened to “ploughing through a haystack.” As one expert revealed in our consultation process: “it may take 1 to 2 weeks for a team of four researchers to complete just the abstract and title screening, assuming full-time focus.” 1

As a result, full systematic reviews, for example, frequently take over a year to complete, with up to 80% of the total effort concentrated in the data-handling stages before any analysis begins. Budgets typically exceed CHF 150,000 per systematic review—placing this type of work out of reach for many organizations, including EdLabs.2 These challenges are aggravated by three major obstacles: 

Fragmented research: Much of the evidence is dispersed across countless websites and databases, often locked behind paywalls or confined to academic silos. Despite valuable resources like the Education Endowment Foundation (EEF)'s Teaching and Learning Toolkit, the International Initiative for Impact Evaluation’s evidence gap maps, and the IES What Works Clearinghouse, there is little consolidation. Each operates in isolation with its own underlying repository of research, further fragmenting the availability and usability of evidence. 

Lack of a shared coding taxonomy: Without a common framework to organize and tag research, it is nearly impossible to systematically compile, compare, or share findings across various contexts and projects. This lack of interoperability weakens collaboration, increases duplication, and hampers data reuse. 

Limited capacity of EdLabs: Many national evidence intermediaries, such as EdLabs, lack the in-house expertise and resources to fully harness available research for producing or commissioning high-quality syntheses. Consequently, they often recreate databases and analytical frameworks from scratch—a process that is both time-consuming and duplicative. Moreover, valuable, pre-coded data remains locked behind paywalls, forcing intermediaries to rely on generic global reports that do not sufficiently account for local contexts, or to produce suboptimal outputs that fail to inform effective policy and practice decisions. 

These challenges culminate in a disjointed evidence ecosystem where duplicated investments and missed opportunities are the norm. In addition, syntheses are infrequently updated and are often produced by external bodies with limited local insight, leaving decision-makers to rely on outdated, generic findings—or, worse, lower-quality, hastily assembled data that increases the risk of bias. Without the integration of coordinated, open-access repositories, living syntheses, and shared standards, the education sector cannot fully capitalize on its wealth of research.

The proposal for this panel will focus on exploring these challenges and presenting an upcoming solution developed by a large Swiss Foundation, aiming to empower EdLabs to deliver more contextually relevant, timely, evidence synthesis of a high quality with the ultimate aim of improving decision-making in education at a country level.

Sub Unit

• General Pool

Chair

• Ines Charro, Jacobs Foundation

Individual Presentations

• Contextualise the Global Evidence Base for Teachers and Policymakers in the Lake Chad Basin. - Patrick Okwen, Effective Basic Services (eBASE) Africa

• Building Better Systems for Evidence with EdLabs - Juan Manuel Hernandez Agramonte Caballero, Innovations for Poverty Action

• The Rise of Knowledge Brokers in Global Education Governance - Samuel Kembou, Jacobs Foundation