Expert Comment: How do we keep children learning when schools have to close?

When schools reopen after the summer, most of us envision classrooms with children sitting with their teachers, learning alongside their peers. But for millions of children around the world, getting to school cannot be taken for granted.

In the Philippines, where our latest research was conducted, schools can be disrupted by typhoons, floods and other natural hazards. And climate-related disruption is becoming increasingly common: UNICEF estimates that at least 242 million students in 85 countries had their schooling disrupted by extreme weather in 2024. For many of us, the experience of COVID-19 made the consequences of school closures painfully familiar: children suddenly lost not only lessons, but the routines, relationships and social world that school provides.

These crises raise a question: how do we keep children learning when they cannot get to school?
 

Our recent research, published in Nature, suggests that relatively simple technology can play a role. Across five countries - India, Kenya, Nepal, Uganda and the Philippines - we found that phone-call tutoring produced meaningful learning gains during periods when children could not attend school. The circumstances varied, but all five countries experienced school disruption during the COVID-19 pandemic, with some facing exceptionally long closures. In the Philippines, typhoon Rai provided an unexpected additional test of whether learning could be protected when another crisis struck.

In Uganda, for example, the proportion of Grade 4 children able to perform division increased from 10% among children in the control group to 48% among those receiving the phone-based intervention. That result matters as much for how we tested the approach as for what it showed in Uganda.

We conducted five randomised controlled trials across five different countries. Randomised trials are the same standard of evidence used to test new medicines, but conducting them across multiple countries is surprisingly rare: less than 1% of impact evaluations are conducted across multiple countries at once. Five trials testing the same question across five diverse settings is therefore an unusually strong evidence base. It allowed us to ask a question that is crucial for policymakers: does an approach continue to work when it moves beyond the conditions of a single study or context?

The answer was encouraging, with one important qualification: it was not the mobile phone alone that made the intervention work.

The approach depended on two things: the platform and the pedagogy. The platform was the mobile phone, which meant that children could be reached even where computers and reliable internet were unavailable. The pedagogy involved targeting instruction to each child's level of learning.

Both pieces matter. Giving children access to technology does not automatically lead to better learning. The technology needs to deliver the right kind of teaching.

There is another lesson here that I think is particularly important for education policy. We tested different delivery models, including programmes delivered through NGOs and through government teachers. An intervention that works under the controlled conditions of an NGO programme may not work in the same way when incorporated into a government education system, where programmes have to reach much larger numbers of children and operate within existing systems.

That is why implementation matters as much as effectiveness. It is not enough to know that something can work. Governments need to know whether it can be delivered, at scale, through the systems they already have.

Our multi-country research also revealed something unexpected: a “learning curve”. The effects of the phone tutorials grew stronger across the sequence of trials, as we got better at targeting instruction to each child's level. In other words, implementation itself improved as we tested, adapted and refined the approach.

This is an important shift in how we think about education research. Too often, research asks: does this intervention work? For policymakers facing a crisis, the more useful questions are: will it work here? Can we deliver it through our education system? And can we get better at delivering it over time?

This is particularly important because school closures are not going away. Conflict, epidemics and extreme weather can disrupt education with little warning. We cannot predict every crisis, but we can make education systems more resilient before one happens, repair the roof as the sun is still shining.

That means building contingency mechanisms while schools are open. Governments can identify how teachers, schools, technology and existing education infrastructure could be mobilised if children suddenly cannot attend school. They should not have to invent a response from scratch in the middle of a crisis.

None of this means that mobile learning should replace schools. Children need teachers, relationships, interaction with other children and the social environment that comes from being physically in school. Remote learning is a fallback for when the normal system is temporarily unavailable. It is not a replacement for the classroom.

As children across the UK return to school this September, it is worth remembering that a smooth start to the school year is not guaranteed everywhere. For many children around the world, disruption is already a reality.

The best time to build resilient education systems is while schools are open. Knowing what works is only half the challenge. Governments have to be able to deliver it at scale, through the systems they already run, before the next crisis closes classrooms.

For more information about this story or republishing this content, please contact [email protected]