Pratham’s learning model tests what children can actually do
Published on 20 August 2026 at 02:02 GMT
By Editorial Team SDG4
School enrolment can conceal a basic problem: children may advance through grades without mastering the reading and arithmetic expected in the early years. Pratham has spent three decades turning that gap into something communities and education systems can measure, discuss and act upon, using short assessments and instruction organised around what children can do rather than the grade listed beside their name.
The organisation began in Mumbai in 1995 and now works across education, youth training and research. Its approach to foundational literacy and numeracy is known as Teaching at the Right Level, or TaRL. The method uses a simple assessment to identify a child’s current reading and arithmetic level, groups children by those levels for part of the day, and uses targeted activities to help them progress. Children are reassessed and regrouped as their skills change.
This is a practical departure from age- and grade-based instruction. A child enrolled in Grade 5 may still be working on recognising words, while another may read a short story. Under TaRL, those differences shape the immediate teaching task. Pratham says its direct Learning Camps usually provide 30 to 50 days of instruction, while government-partnership versions may run for four to six months within the school year. Staff, teachers and community volunteers can all be involved, depending on the programme.
Measuring the learning gap
The wider context comes from the Annual Status of Education Report, or ASER, a rural household survey facilitated by ASER Centre, the research and assessment unit of Pratham Education Foundation. ASER 2024 reached 649,491 children in 17,997 villages across 605 rural districts. Local organisations or institutions conducted the survey in each participating district, giving the exercise a citizen-led structure as well as national scale.
The latest results show both recovery and a persistent shortfall. Among all rural pupils in Grade 5, 48.8 per cent could read a Grade 2-level text in 2024, up from 42.8 per cent in 2022 but still below the 50.5 per cent recorded in 2018. In arithmetic, 30.7 per cent of Grade 5 pupils could solve a division problem in 2024, compared with 25.6 per cent in 2022 and 27.9 per cent in 2018. Enrolment among children aged six to 14 remained high at 98.1 per cent.
These figures do not measure every part of education, and ASER covers rural households rather than the entire school system. They nevertheless separate attendance from learning outcomes and expose large differences within the same grade. The assessment tools make the distinction accessible beyond specialist testing. The tools use recognisable tasks, such as letters, words, short texts, number recognition, subtraction and division, so results can be explained to families and local volunteers without relying on complex scores.
Evidence, adaptation and scale
The approach has also been tested through randomised evaluations. The Abdul Latif Jameel Poverty Action Lab, or J-PAL, reports that six evaluations across seven states found test-score gains ranging from 0.07 to 0.70 standard deviations. Its review concludes that targeted instruction improves basic reading and mathematics when programmes align teaching with learning level. The evidence also shows that a method’s design is not the same as reliable implementation: grouping, dedicated teaching time, suitable activities, monitoring and support all affect results.
That implementation question becomes sharper when a programme moves from camps led by an NGO to government classrooms. Pratham describes two main routes: direct delivery with communities and schools, and partnerships in which government teachers use the method with training and mentoring. Its elementary direct programmes operated across 21 states and union territories and reached more than 347,000 children in 2024-25, according to the organisation. This is organisational reporting rather than an independent estimate of learning impact, and scale alone does not establish effectiveness.
International expansion adds another layer. J-PAL reports that TaRL has reached more than 80 million children across India and Africa, with governments and non-profits adapting the model to different languages, curricula and administrative systems. The model follows a replicable sequence: assess, group, teach, monitor and regroup. Whether that sequence retains its quality at scale depends on training, time, supervision and the capacity of local systems.
Community participation is central to both ASER and parts of Pratham’s teaching model. Volunteers make household assessment possible, while mothers, local groups and schools feature in community programmes such as Hamara Gaon. This can make learning levels visible outside the classroom, but it also raises practical questions about volunteer continuity, local capacity and the relationship between community action and the state’s responsibility for public education. Pratham’s own materials present community participation as a complement to school and government partnerships, not a substitute for them.
The work connects directly to SDG 4 (quality education), which includes access to education and effective learning outcomes. The ASER figures show that high enrolment can coexist with many pupils lacking specific skills associated with earlier grades.
The central tension is now less about whether foundational learning can be measured simply than about how evidence changes everyday instruction. Pratham has helped establish a model in which a short assessment is linked to an immediate teaching response. The remaining question is whether public systems and community partnerships can preserve that link consistently, while recognising the limits of household surveys, programme data and models adapted across very different settings.
Further information:
• ASER Centre, ASER 2024 National findings, supporting the survey coverage, enrolment and national reading and arithmetic figures.
• Pratham, Teaching at the Right Level, describing the assessment, grouping, Learning Camp and government-partnership models.
• Pratham, Elementary Years, supporting the two implementation modes and the reported reach of direct programmes in 2024-25.
• Abdul Latif Jameel Poverty Action Lab, Tailoring instruction to students’ learning levels to increase learning, supporting the randomised-evaluation findings.
• Abdul Latif Jameel Poverty Action Lab, Teaching at the Right Level to accelerate learning, supporting the account of international adaptation and reach.
