Learning to read is foundational to education, opportunity, and long-term development. Despite decades of investment in education, millions of children still do not acquire this most basic skill. A major new synthesis of global research makes one thing clear: the problem is not whether children can learn to read, but whether they are being taught in ways that align with evidence.
According to Effective Reading Instruction in Low- and Middle-Income Countries: What the Evidence Shows, endorsed by the Global Education Evidence Advisory Panel, 70% of children in low- and middle-income countries (LMICs) cannot read and understand a simple, age-appropriate text by age ten. This figure reflects a sharp deterioration from pre-pandemic levels, when 57% of children were affected, a condition referred to as learning poverty.
Access has improved, learning has not
The challenge is not school access. Globally, primary school enrolment in LMICs increased from 62% in 1970 to 89% in 2023. Yet expanded access has not translated into effective learning.
A large-scale analysis of more than 500 000 students across 48 LMICs and 96 languages found that, after three years of schooling, more than 90% of learners could not identify letter names or letter sounds, or read simple words at expected levels. This failure persists despite LMICs spending, on average, about 4% of GDP on education.
Why children are not learning to read
The evidence points to a central cause: instructional approaches that are not aligned with research on how children learn to read.
The paper identifies widespread use of methods that delay or underemphasise decoding, rely on memorisation or repetition, and expect children to develop reading skills on their own. Observational studies across LMIC classrooms document practices such as rote copying, repetitive choral reading, and minimal explanation – approaches shown to be ineffective at developing skilled readers.
Drawing on more than 120 studies conducted across Africa, Asia, Latin America and the Caribbean, and the Middle East, covering over 170 languages, the report synthesises a clear set of principles grounded in the Science of Reading.
The evidence shows that reading with comprehension depends on two essential domains:
- Decoding: the ability to translate written symbols into sounds
- Language comprehension: the ability to understand the meaning of words and texts
To develop these abilities, children need explicit, systematic instruction in six core skill areas:
- Oral language development, including vocabulary and listening skills
- Phonological awareness, or the ability to identify and manipulate sounds in spoken language
- Systematic phonics instruction, teaching consistent letter–sound relationships
- Reading fluency, enabling accurate and automatic reading
- Reading comprehension, supported by explicit strategies and background knowledge
- Writing, including letter formation, spelling, and composing texts
The paper emphasises that weakness in any one of these areas can prevent children from becoming skilled readers. Instruction must therefore be comprehensive, not selective.
The language of instruction matters
The paper highlights another critical factor: over one-third of children in LMICs are taught to read in a language they do not speak or understand well. Research consistently shows that children learn to read most effectively when instruction begins in their home language. Where this is not possible, significantly more time and support are required to develop oral proficiency before reading instruction can succeed.
Beyond learning outcomes, the paper notes that structured pedagogy programmes incorporating evidence-based reading instruction are likely among the most cost-effective education interventions available. Early investment reduces the need for remediation, lowers repetition and dropout rates, and improves overall system efficiency.
The report concludes that dramatic improvements in reading outcomes are achievable within reasonable timeframes – if education systems act on the evidence. It urges policymakers to:
- Commit nationally to evidence-based reading instruction
- Choose appropriate languages of instruction
- Ensure explicit, systematic teaching across all six core reading skills
- Adapt instruction to language characteristics without abandoning core principles
- Invest in high-quality implementation and teacher support
The global literacy crisis is not a mysterious problem. The evidence on how to teach reading effectively exists, is robust, and spans diverse low- and middle-income contexts. What remains is the political and institutional will to apply it.
