What Works Hub: Scaling Education Reform Globally

Anders Holm, from a foundation in Denmark, stated that the organization is applying the 'What Works' hub approach to biodiversity and rewilding initiatives. This expansion reflects a broader effort to use evidence-based inquiry in diverse fields.

Holm admitted to 'stealing' the hub's core idea, emphasizing its potential to improve foundational practice. The foundation now actively encourages other funders to adopt similar evidence-based methodologies across their portfolios.

First, evidence must challenge funder decisions, especially when intentions are perceived as 'good' but outcomes are unverified.

Second, funders must intensify efforts to make evidence actionable within government systems. This involves translating research into practical steps that can be implemented effectively at scale.

Third, donors must apply the same rigorous evaluation standards to their own operations and strategies as they demand from their partners, ensuring accountability across the board. This self-scrutiny is essential for genuine impact.

This flexibility is crucial for fostering innovation and creating a responsive market.

Data from 11 Randomized Controlled Trials (RCTs) in Ghana and Sierra Leone demonstrated positive overall impacts on literacy and numeracy, validating the OBF approach. Payments for service providers are contingent on independently verified student outcomes, directly linking funding to measurable results.

This model allows for dynamic adaptation to local needs, diverging from traditional input-based funding. It incentivizes providers to find the most effective methods to achieve specified learning goals.

Ahmed Abdullahi Muhammad presented the successful piloting of two models of Teaching at the Right Level (TaRL) in Somalia. The initiative targeted literacy and numeracy across 65 primary schools and 99 internally displaced persons (IDP) settings.

Both settings showed significant improvements, with a remarkable 92% improvement in learning outcomes, often exceeding expectations. This statistical significance highlights the intervention's efficacy in challenging environments.

A key adaptation involved shifting from school-based teacher mentors to utilizing quality assurance officers as coaches. This change streamlined implementation and leveraged existing government structures, contributing to the program's success.

Emily Burkho presented findings on the impact of non-stereotypical stories on student beliefs in Botswana, using AB testing to isolate the effect of empowering content in curricula. The study revealed that tweaking story content significantly shifts beliefs.

In urban areas, the percentage of girls who believed they could not pursue any job they desired decreased by 7.8 percentage points. This indicates a measurable change in self-perception and ambition. Significant shifts in career aspirations towards professions like doctors and teachers were also observed among girls.

However, the results were less impactful in rural areas, primarily attributed to a lack of job prevalence and opportunities. This suggests that while empowering content can change beliefs, external factors like job availability also play a crucial role in shaping aspirations.

Niha Rail discussed the phased handover of supervision, coaching, and data collection responsibilities from the Luminos Fund to the Ethiopian government. The long-term goal was always to transition ownership, recognizing that it is the government's program, with external partners providing support.

The program scaled rapidly, expanding from 135 to 700 classrooms over two years. Despite this rapid expansion, reading fluency was successfully maintained at 23-24 correct words per minute, demonstrating effective scaling without compromising quality.

Success in such transitions hinges critically on strong government championship and sustained investment in supervisor capacity. This ensures that local systems are equipped to manage and sustain educational interventions independently.

An example cited was a 0.64 standard deviation (SD) improvement, which equated to only 10 letter sounds and two words per minute in raw terms.

Standardized effect sizes are not universally comparable across different languages, grade levels, or assessment instruments. This variability complicates the interpretation of progress and makes it difficult to draw meaningful comparisons.

The lack of raw effect reporting makes it challenging for practitioners to interpret the true educational progress of students. Transparent reporting of raw gains alongside standardized metrics is essential for practical application.

This discrepancy can lead to different conclusions about which interventions are most effective.

Many studies fail to report essential raw data or provide sufficient information on instrument reliability, which complicates rigorous meta-analyses and limits the utility of findings. Without these details, comparing and synthesizing research becomes problematic.

Moving forward, transparency in reporting raw gains is as crucial as standardized metrics. Providing both types of data allows practitioners and policymakers to make more informed decisions based on a complete picture of impact.

Dr. Akashi Kung stated that the primary bottleneck in India's education reform is not the generation of data, as much data has been collected. Instead, the challenge lies in establishing a reliable path from evidence to action, translating research into implementable policies.

Programs intended for state-level scale require different evidence strategies compared to small-scale pilots. What works in a controlled pilot may not translate effectively to a vast, complex governmental system without tailored approaches.

Government ownership of data is identified as a critical factor for policy uptake. When governments actively collect, analyze, and interpret data, they are more likely to integrate evidence into their policy-making processes, leading to more sustainable reforms.

Mumi highlighted that South Africa developed distinct benchmarks for its 11 official languages, avoiding the common mistake of treating African languages as monolithic. This approach acknowledges the unique linguistic characteristics and educational needs of each language group.

Language family differences, such as variations in word count required to convey the same meaning, necessitate tailored benchmarks. A one-size-fits-all approach would inaccurately assess proficiency and progress across diverse linguistic contexts.

Large-scale assessment results in South Africa revealed significant gender and socioeconomic status gaps in learning outcomes. This underscores the importance of disaggregated data to identify disparities and inform targeted interventions.

When implemented with high fidelity, meaning teachers accurately group children by level, the impact doubles.

Government-delivered programs typically show a 20 percentage point gain. However, when Protham staff deliver the program, the gains are double, indicating the importance of expert implementation.

High-fidelity implementation, particularly through accurate grouping by level, can increase the impact to 0.85 standard deviations. This emphasizes that not just the program, but how it is delivered, critically determines its effectiveness.

Bumi described South Africa's policy shift, triggered by PIRLS (Progress in International Reading Literacy Study) results showing 81% of grade four learners could not read for meaning. This prompted a national focus on early literacy, moving beyond solely concentrating on school-leaving examinations.

The focus shifted from grade four to intervening as early as grade one, recognizing the critical importance of foundational literacy. This reorientation aims to address learning deficits at their earliest stages.

The strategy also includes leveraging local languages in instruction and building in-country research capacity. This dual approach ensures culturally relevant education and strengthens local expertise in evidence-based reform.

This constant negotiation plays out daily, resembling a complex game.

A key practical limitation is the inability to spend 1.5 hours assessing a single child at scale due to logistical and resource constraints. This necessitates finding efficient yet effective assessment methods.

The process involves balancing ideal academic competencies with government mandates and logistical realities. Aligning assessment directly with programmatic intervention is crucial to ensure that measures genuinely reflect the impact of the educational efforts.

Dr. Samuel Kimu critiqued the common assumption that research 'push' automatically creates 'pull' from policymakers, implying that investing in research alone guarantees uptake. This perspective often overlooks the complexities of policy implementation.

The 'labification' of education policy, which equates the form of research with its function in policy, was also criticized. This approach often neglects the practical realities and administrative capacities required for successful integration.

There is a narrow focus on program adoption rather than a comprehensive understanding of public administration and delivery. Effective policy implementation requires attention to systemic factors beyond just the merits of a specific program.

Akash shared his experience of establishing an 'ed lab' within a stringent six-month deadline, a target he set for himself before an inevitable administrative transfer. He commenced work in August and successfully inaugurated the lab by January 2026.

He applied the Pareto principle, focusing intervention efforts on the 20% of schools responsible for 80% of failures. This strategic approach maximized impact within limited time and resources. The lab was formally inaugurated on India's Republic Day, a symbolic gesture.

Akash strategically framed the 'ed lab' project as a direct means to improve 10th and 12th board exam results, which appealed to both public and political interests. This objective aligned with short-term governmental goals, ensuring crucial buy-in.

The project quickly yielded tangible results, leading to a jump to 3rd position in the state for 10th-grade board exam results and 5th position for 12th-grade results in terms of improvement. These demonstrable successes solidified support for the initiative.

By linking the ed lab's efforts directly to widely recognized performance metrics, Akash secured essential backing. This approach proved effective in a system where immediate, measurable outcomes often drive political and public enthusiasm.

Hamad Tarik emphasized the importance of identifying 'binding constraints' when working with governments, rather than attempting to fix every issue simultaneously. This approach prevents stakeholders from being 'bogged down' by the entirety of systemic failures.

Using granular data enables real, specific conversations with government partners, moving beyond generalizations to address precise problems. This data-driven dialogue fosters more effective collaboration and targeted solutions.

Institutionalization through documentation, such as meeting minutes and official notifications, is crucial for embedding reforms within government systems. This formalizes changes and ensures their sustainability beyond individual tenures.

In The Gambia, stakeholders showed overwhelmingly positive attitudes toward using local languages for instruction, despite English maintaining its status as a language of prestige. This shift was largely driven by community input, addressing concerns about children's learning and teaching effectiveness.

Language mapping facilitated the strategic integration of seven major Gambian languages into classroom instruction. This inclusive approach recognized the linguistic diversity of the nation and aimed to improve foundational learning outcomes.

The stakeholder-driven policy culminated in the Cabinet's approval of the Language and Education Policy 2025-2030 in March 2026. This formal endorsement signifies a significant shift towards culturally and linguistically appropriate education.

Permod Kumar highlighted a systemic issue: foundations often receive credit for successes, while governments are held accountable for systemic failures. This dynamic impacts how interventions are presented and perceived.

He emphasized that one-size-fits-all 'imported' programs frequently fail in local contexts. Effective reform requires adapting global strategies to fit local needs, including appropriate tools and context-specific implementation methods.

Success demands moving beyond abstract research metrics like T-values and P-values to simple, actionable data that teachers can use directly in the classroom. This focus on practical utility ensures that evidence translates into real improvements in learning.

This perception discourages teachers from fully utilizing their capabilities, even when they know they can achieve more.

This lack of systemic support for high performers means that individual efforts are not adequately recognized or rewarded. Instead, excellence is often met with additional burdens rather than opportunities for growth or advancement.

Teachers frequently feel overburdened with administrative tasks, further contributing to their reluctance to take on more responsibilities. Addressing these systemic issues is crucial to fostering a motivating environment for educators.

This approach promotes shared responsibility and broader impact.

A long-term focus is essential, as quick fixes are often unsustainable and fail to address deep-rooted systemic issues. True reform requires sustained commitment and collaborative strategies over extended periods.

The discussion highlighted the importance of transitioning from research findings and Randomized Controlled Trials (RCTs) to actual classroom implementation. Bridging this gap requires concerted efforts from all stakeholders to translate evidence into practical educational improvements.

Ranil pointed out the stubborn challenge of a significant disconnect between policy and evidence-based spending in global education. Despite mounting evidence on the high value and importance of foundational learning, spending profiles exhibit considerable inertia.

There is excessive inertia in existing spending profiles, making it difficult to reallocate funds effectively. This means that even as new evidence emerges, established funding patterns persist, hindering adaptive investment.

Consequently, there is persistent difficulty in shifting funds from tertiary education to foundational learning, where the evidence suggests the greatest impact can be achieved. Realigning these priorities is critical for more effective aid.

A questioner highlighted that the UK aid budget has been drastically cut, with the education aid budget experiencing an even more severe reduction. This substantial cut impacts global education initiatives.

The UK's aid budget decreased significantly from 0.7% to 0.3% of Gross National Income (GNI), leading to widespread implications for various sectors. Within this shrinking total, protecting education spending has become increasingly difficult.

A strategic shift in aid allocation towards fragile and conflict-affected states is also underway. While this aims to address urgent needs, it necessitates difficult decisions about funding priorities across the broader education landscape.