Age or Level? The Hard Choice When a Refugee Child Returns to School

What happens to a 12-year-old in a class of 7-year-olds, and what research says works.

In 2017, UNHCR told the story of a Rohingya girl in Kuala Lumpur who started school at 14, because there had been no school for her before.¹ She sat in a primary-level class with younger children, learning English, maths and science. She felt self-conscious about being older than her classmates. Her mother, who never went to school herself, kept telling her the same thing: this is for you.

Her story stayed with me, because it holds a question that I had never thought about. When a child has missed years of school, where do you put them? With children their age, or with children at their level? I went looking for what research says. Here is what I learned.

Why is this such a common question?

So many refugee children lose years of school. UNHCR estimates that 5.8 million refugee children were out of school in 2025, 44 per cent of all school-aged refugee children.² Many more had their schooling interrupted along the way: in 2016, UNHCR estimated that refugee children arriving in a new country had often missed three to four years of school.³

And the path narrows as children grow older.

One in three refugee children reaches secondary school

Gross enrolment rate of refugee children by level of education, 2025

Primary

65%

Secondary

32%← the big drop

Higher education

11%

Even a child’s age is not always clear. A teacher of refugee students in Canada described how five of the nineteen children in her class shared 1 January as their birthday: a date assigned on paper, not the day they were born.⁴

So every school that welcomes refugee children has to make this choice, often with little to go on.

What happens when you place a child by age?

The child stays with peers. That matters: friendships, belonging and not feeling like the odd one out are part of learning too.

But the lessons assume years of schooling the child has not had. Teachers are under pressure to cover the curriculum for their grade, and research from J-PAL notes that they tend to teach to the top of the class.⁵ A child who is three years behind can sit in every lesson and still fall further behind.

What happens when you place a child by level?

The lessons now fit. But the child is suddenly the oldest in the room, sometimes by many years.

Researchers from NYU’s Global TIES for Children studied 448 Syrian refugee children in Lebanese public schools.⁶ The most consistent warning sign for later difficulties was being older than expected for the grade.

Working memory is what helps a child focus, follow instructions and manage emotions. An honest caveat: the study cannot fully separate the effect of the placement itself from what came before. Being older than your grade is often a sign of a hard road behind you. But it tells schools to watch these children closely, and to support more than their academics.

Is there a better way?

Yes, and this is the encouraging part. The best evidence I found does not choose between age and level. It combines them.

Teaching at the Right Level. In this approach, developed by the Indian organisation Pratham and tested by researchers at J-PAL, children stay in their regular class, but for part of the day they are regrouped by what they can actually do, not by age or grade. Instructors check progress regularly and move children up as they improve.

Six randomised evaluations in seven Indian states found gains in test scores, and, according to J-PAL’s summary, the children who were furthest behind gained the most.⁵ In one version, short learning camps of 40 to 60 days during the school day doubled the share of children who could read a paragraph or story, from 24 to 49 per cent.

There is one condition. In an evaluation where teachers kept teaching the regular curriculum instead, the approach had no effect. It works best when there is dedicated time for it.

Newer research backs this up, with some nuance. In 2023, the Global Education Evidence Advisory Panel, an independent group of experts co-hosted by the World Bank, UNICEF, USAID and the UK government, named teaching by learning level rather than grade one of three “great buys” in education.⁷ In Zambia, a large randomised trial of the government’s own programme found about two to three months of extra learning over two years.⁸ But a recent review of the rigorous studies found mixed results: programmes run by NGOs or closely supervised facilitators tended to do better than those run by teachers within the regular system.⁹

Accelerated education. For children who are too far behind or too old for a regular class, accelerated education programmes offer a condensed, age-appropriate route, so a 13-year-old does not have to sit with 6-year-olds for years.¹⁰ An inter-agency working group set up by UNHCR has written ten principles for doing this well.¹¹ The evidence base is still growing.

Support beyond the lessons. The NYU researchers add that children who are catching up also need social and emotional support that fits their age and their level.¹²

What does this mean in Malaysia?

In Malaysia, refugee children cannot attend public schools at all.¹³ They learn in community learning centres, many of them built and run by refugee communities themselves. Most of these centres teach at primary level, and only a few dozen offer secondary education.¹⁴

That means teachers there face this question every day: a 14-year-old who has never been to school, children of very different ages in one class, very few resources. Funding cuts in 2026 have made it harder still.¹⁵

And catching up is possible. In the same 2016 report, UNHCR told the story of Nawa, a Somali refugee who first went to school at 16, in a community learning centre in Malaysia.³ Less than four years later, she was about to start a university foundation course, while volunteering as a teacher at her old school.

The research suggests that much of what works is not expensive. Simple assessments, grouping by level for part of the day, and time to do it well. That is exactly the kind of thing that can be supported.

What you can do

The teachers and communities running these schools are already doing the work. The most useful thing is to strengthen what they do.

Or share this piece with a teacher, especially one who welcomes newcomer children.

Read further

About images and names. We do not publish photos in which children can be identified without informed consent from them and their guardians, and we never share the locations of learning centres. When we tell individual stories, we may change names to protect people.

Sources15 sources

1. UNHCR (2017). Rohingya student beats the odds to get schooling in Malaysia. ↩
2. UNHCR (2026). Education Report 2026: Pathways to the Future. Out-of-school figures and gross enrolment rates for 2025. ↩
3. UNHCR (2016). Missing Out: Refugee Education in Crisis. Years of school missed; the story of Nawa. ↩
4. Fitzpatrick (2018). Grade level placement for refugee students. Morning Watch, Memorial University. ↩
5. J-PAL. Tailoring instruction to students’ learning levels to increase learning. Policy insight on Teaching at the Right Level. ↩
6. INEE / NYU Global TIES for Children. Disruptions in schooling: lessons from the refugee experience in Lebanon. ↩
7. World Bank / Global Education Evidence Advisory Panel (2023). 2023 Cost-Effective Approaches to Improve Global Learning (“Smart Buys”). ↩
8. J-PAL. Teaching at the Right Level in Zambia. ↩
9. AFLEARN. The limits of a “great buy”: a review of Teaching at the Right Level in Africa. ↩
10. UNHCR. Accelerated Education Working Group. ↩
11. INEE. Accelerated Education Working Group and the ten principles for effective practice. ↩
12. NYU Global TIES for Children. Policy brief on Syrian refugee children in Lebanese public schools. ↩
13. UNHCR Help Malaysia. Education. ↩
14. Universiti Malaya, JATI journal. Refugee education in Malaysia. ↩
15. UNHCR Malaysia (March 2026). Update on refugee education and funding cuts. ↩