BLOG POST

A New Research Boost for Education: Launch of the What Works Hub for Global Education

An important new program to support education research and policy launched in September. The “What Works Hub for Global Education (WWHGE) is established in Oxford’s Blavatnik School of Government and funded by the UK’s Foreign Commonwealth and Development Office. It will operate over six years with an initial budget of £55 million (or US$70 million). The Blavatnik School was also host for the pathbreaking, eight-year 40 million pound (US$50 million) RISE program (Research on Improving Systems of Education), which closed this year.  

Several things connect the RISE and WWHGE programs. First, the UK has once again made a bold commitment to the kind of substantial long-term funding that permits serious advances in global knowledge. Second, a launch conference in September generated excitement among researchers and development partners to be part of a network sharing cutting edge research experiences. The September conference attracted a “who’s who” of global education, including innovators, implementers, researchers and funders. 

At the conference, Academic Director Noam Angrist described the core goal of the program as advancing “implementation science”. The research question is not so much to measure the impact of single interventions in controlled conditions (“what works” in a randomized trial) as to understand how and why such programs achieve (or fail to achieve) desired results when operating at scale in different countries’ real-world conditions. Rachel Hinton, the FCDO intellectual architect of the WWHGE and RISE programs, presented an example from her own anthropological research investigating why a child feeding program in Kenya failed to increase children’s weight. The answer was a program design that transferred a large part of the food to parents, who could not always prioritize feeding their children. Assuring that the children were fed during the program’s hours of operation could have avoided this. She also presented a graphic picture of the implementation complexity that government officials and their development partners often face. 

To increase the impact of the WWHGE, the research menu is concentrated both thematically and geographically. The theme is programs aimed at increasing foundational literacy and numeracy. The geographic ambits are India, Pakistan, Tanzania and Rwanda. Note: the Hub is also channeling resources to a broader set of African countries (Botswana, Ghana, Sierra Leone, Kenya, Nigeria, and South Africa) and Bangladesh in South Asia.  

The first three were also a focus of RISE research which prompted me to realize that in 2014 researchers had to build priorities, plans and create local teams and international research partnerships from scratch. Now, in the cases of India, Pakistan and Tanzania there are established country research teams (in every case led by local researchers in collaboration with international researchers) and an important amount of high-quality RISE-funded research to build on, ranging from teacher financial incentives to hiring standards and the politics of reform adoption and implementation.  

Comprehensive insights from the research papers presented at the conference and insights from the excellent panel discussions can be found in these real-time reports from Clio Dintilhac, and blog posts from Natasha Ahuja and Pedro Freitas

What follows here is just a smattering of the new research strategies for the focus countries: 

In India, Abhijit Singh described how the team’s work will focus on analyzing implementation of India’ s 2020 New Education Plan, which sets the goal of universal foundational literacy and numeracy (FLN). Researchers will study how the Plan is interpreted and implemented in different states and between government and non-government schools. Since the NEP for the first time incorporates pre-school into FLN, there is wide scope for adding to an almost nonexistent research base on pre-primary schooling in India looking at how pre-school organization, staffing, curriculum, activities and implementation quality affect foundational learning in the early primary grades.  

In Pakistan, Tahir Andrabi described research on a “targeted instruction program (TIP)” being analyzed in two districts with three different modes of teacher support: paper materials; optional paper or tablet-base materials or mandatory tablet-based materials. After noting the high level of teacher rotations across schools in both districts, the team has extended the research to understand what drives teacher rotation and its impacts on the sending and receiving schools.  

In Tanzania, in the context of high-level government policy demand to improve instructional quality, researchers want to examine the comparative effectiveness of school vs. cluster-based teacher continuous professional development activities. With the goal of getting teachers to learn together and practice new teaching methods, in many countries the most efficient strategy is to promote teacher interaction at the school level, creating “professional learning communities” (Canada) or “lesson study” circles (Japan) or explicit time in the school week for teacher collaboration (Finland). In contexts where rural schools are small and mentors or other sources of support are limited, Colombia’s “microcentros”, with roving mentors and Chile’s focus on mentoring novice principals can be alternatives. A question in the Tanzanian context is how well these models can work, and if leveraging local district offices to supply mentors increases the likelihood of productive school-level interaction. Researchers will build evidence on the cost-effectiveness of alternative approaches.

An example of how the WWHGE will fund particularly innovative research on a wider set of African countries is the comparative study of the operation of “outcomes funds” on implementation and learning results in Ghana and Sierra Leone that Dr. Clare Leaver presented. Outcomes funds transfer resources to providers based on their learning results. In the first phase, school districts will be rewarded for improvements in learning that exceed those of control districts. There is much to learn about how schools and districts respond to results-based performance incentives (what they change, what they prioritize, and the variance in strategies, implementation and results across different schools.) Also promising is the chance to compare results from two countries operating programs with the same core design features. 

The addition of Rwanda as a focus country is exciting, as the government in the past few years has begun devoting significant resources to measurement of learning outcomes. It is currently the only country in Africa, (or for that matter anywhere in the global south) producing annual data on student learning at the level of individual classrooms. This creates scope to do really pathbreaking work on teacher quality – for example, measuring how much teachers’ effectiveness varies across classrooms, schools, and districts. As Dr. Leaver, co-lead of the Rwanda research team said: “Rwanda is pushing the envelope of education policy, which will have great benefits not only for its own students, but for students across the developing world.” 

It was classroom-level data like this which in the US led to breakthrough studies by Stanford’s Eric Hanushek, revealing tremendous variance in learning outcomes at the classroom level, even in the same grade in the same school – differences that could represent up to one half of a year of schooling. Harvard’s Raj Chetty with time series data on classroom level learning outcomes over several years of primary school was able to follow individual children’s educational trajectory. Children exposed to “highly effective teachers (in the top 20 percent of the distribution in terms of their students’ learning gains) not only learned more in the short term, but had better long-term life outcomes (college attendance, employment and incomes).  

The within-school variance in teacher practice across classrooms has also been documented through standardized classroom observations in Latin America, but researchers there have never been able to link it to learning outcomes for lack of annual, classroom level learning data. With the benefit of Rwanda’s classroom level learning data, it will be fascinating—and truly pathbreaking—to explore how hiring standards, incentives, training programs, curriculum reforms and other early grade interventions affect teachers’ classroom effectiveness, how quickly, and the impact on foundational literacy and numeracy. 

One of the biggest challenges for maximizing the policy impact of the WWHGE is the same constraint faced by the RISE program research (and indeed, by all other cross-country research programs). It is the lack of comparable data in three critical areas: costs, learning outcomes, and classroom interaction between teachers and students. In some sense, harmonizing cost data may be the least challenging. JPAL and the B2 initiative have already made substantial progress on this. But for a reality check on my assertion here, see a great new piece by Christine Beggs, Clio Dintilhac and Michelle Kaffenberger.  

The biggest, and essential, challenge is collecting comparable learning data. The Hub group is well aware of the challenges in learning measurement. As Dhir Jhingran of the Language and Learning foundation reminded the conference: “Data is political” and “failure is costly”. Countries are often reluctant to appear worse than their neighbors. On the other hand, in Latin America, the regional assessment ERCE and participating in PISA have had demonstrable impact in stimulating reform.  

Things may be changing in Africa. The technically well-regarded PASEC regional test has expanded beyond francophone countries to include lusophone countries and now Nigeria as well. As PASEC includes a second-grade oral assessment of foundational literacy and numeracy, it could be a very relevant and helpful tool for the Hub’s research measuring and understanding progress. 

Equally valuable is comparable classroom observation research. Virtually every program or reform in education can only affect student learning by somehow changing the interaction of teachers and students in the classroom. The introduction of scripted learning materials for teachers and the use of student groupings to facilitate “teaching at the right level” are perfect examples. The more we can closely measure teachers’ uptake and facility with new approaches and supports, using a common classroom observation instrument, the more scope there is for truly important, truly comparable research findings and implementation insights. Good examples are research highlighted at the conference from Haiti (Understanding variation in early grade reading program effects to drive improvement) and South Africa focused on teachers’ uptake of new foundational learning programs. In the South Africa case, researchers were able to use data from tablet-based programs to monitor teachers’ implementation.  

To promote “implementation science” of how programs work at the classroom level, the Hub is also bringing together the relatively small set of researchers interested in classroom observation for joint work on one or two instruments that embed all the best available elements—that have proved correlated with learning outcomes. The goal thereafter would be to encourage all WWHGE funded researchers to draw on this (or these) instruments to conduct parallel classroom observations.  

In a world where most development assistance has scattered priorities, small budgets and often short-term horizons, the WWHGE, and RISE before it, are extremely important. They are practically unique in delivering the resource envelope and timeframe that serious research requires. Equally importantly, they “crowd in” resources from other donors who might not have the political or financial ability to be as bold and strategic as FCDO. Over its lifetime, RISE attracted support increased support from the FCDO and the Gates Foundation and Australia and was able to increase the number of countries in the program and launch additional work around common themes. WWHGE is already benefiting from $2 million in support from the Gates Foundation and relationships with the World Bank, the Global Partnership for Education, USAID, UNESCO-IIEP, the British Council, Building Evidence in Education (BE2), the Global Education Evidence Advisory Panel (GEEAP), the Learning Generation Initiative, UNICEF Innocenti, and the Jacobs Foundation. This set of partners represents an overall budget for WWHGE implementation research on foundational learning on the order of US$ 70 million. Hopefully, the quality of Hub research and its direct applicability to countries’ education policies will attract sustained funding over time from these and the many additional global partners committed to foundational learning.  

The What Works Hub for Global Education is off to an exciting start. Thanks to the pathbreaking work of NGOs such as Pratham, Room to Read and research in Kenya by RTI, as the World Bank’s Luis Benveniste observed: “We know a lot about “what works”; now it is time to get on with implementing it.” In precisely this context, research supported by the Hub—that actually goes beyond “what works” to give us insight into “how” and “why” things work—will be crucial for faster progress. 

Disclaimer

CGD blog posts reflect the views of the authors, drawing on prior research and experience in their areas of expertise. CGD is a nonpartisan, independent organization and does not take institutional positions.


Image credit for social media/web: Media Lens King / Adobe Stock