"Over the last ten years, ASER's headline has barely changed. Only about half of all Grade 5 children in rural India can read a simple Grade 2-level text. We have solved the attendance problem. We have not solved the learning problem." — Rukmini Banerji and Esther Duflo, 2015
In a primary school in a tribal block in Koraput, a Class 4 student cannot read the words on the page in front of her. She has attended school regularly for four years. She has not been failed — India's education system promotes students automatically regardless of whether they have learned what the previous year required. She sits in a Class 4 classroom being taught Class 4 material, and she understands almost none of it. By Class 7, she will likely drop out. The system will record this as a dropout problem. It is, in fact, a learning problem that became a dropout problem because nobody caught it at Class 1.
This is Odisha's central education challenge. And it is not unique to Odisha — it is India's defining education paradox: a generation of children who are in school and not learning.
Odisha has, by any measure, made remarkable progress on access. The Gross Enrolment Ratio at primary level is near-universal. Government investment in school infrastructure, teacher training, residential schools for tribal students, midday meals, and free uniforms has been substantial and documented. The High School Certificate pass percentage rose from 69.6% in 2005 to 94.2% in 2022–23 — a genuine achievement. Tribal female literacy grew by 76.29% over a decade, against a national average of 41.97%.
But the PARAKH 2024 national assessment — the most comprehensive recent measurement of foundational learning across Indian states — placed Odisha among the lowest-scoring states in language proficiency, alongside Bihar, Jharkhand, Madhya Pradesh, and Rajasthan. And Odisha's overall dropout rate increased from 9.1% in 2023–24 to 9.8% in 2024–25, bucking national declines. At secondary level, Tribal boys face a dropout rate of 35.3%.
The access battle is largely won. The learning battle is not. This Primer is about the latter.
Part One: The Learning Landscape — What the Data Actually Shows
Odisha's current literacy rate for persons aged 7 and above stands at 73.6% overall — 82.2% for males and 64.8% for females — against a national average of approximately 80.9%. The rural figure is 70.8%, against 85.8% urban. These are aggregate numbers. In tribal districts, in remote blocks, in communities where the nearest government school is understaffed and the teacher speaks a language different from the children, the numbers are substantially worse.
ASER 2023 found that only about 57% of Class 5 students in rural India can read a Class 2-level text fluently — a figure that represents a decade of near-stagnation in learning outcomes despite enrollment growth. For tribal students in blocks with lower teacher attendance and higher linguistic diversity, the equivalent figure is lower still.
A 2023 audit of 7,146 schools across 98 blocks in Odisha, conducted by the Atma Shakti Trust, found that 73.85% did not adhere to RTE norms for classroom adequacy. Many were operating in makeshift or overcrowded spaces. Basic sanitation and water access remained inconsistent.
The dropout architecture reveals where learning failure converts into system exit. Overall primary dropout is low — gains here are real. Upper primary (Classes 6–8) dropout stands at 2%. But at secondary level, the rate rises sharply — to around 12.8% overall, with tribal boys at 35.3% and tribal secondary dropout historically above 30%. The pattern is consistent with what educational researchers call the "learning poverty trap": children who exit primary school without foundational literacy and numeracy are structurally set up to fail at secondary level. The dropout at Class 9–10 is often the lagged consequence of learning failure at Class 2–3.
Odisha's government has correctly identified this dynamic. The state's FLN programme — implemented through SCERT Odisha under the national NIPUN Bharat mission — has developed tribal mother-tongue materials, trained teachers through cluster-level meetings, and introduced the SAMHATI multilingual education framework for children who enter school speaking languages other than Odia. These are the right interventions. The challenge, which audit reports and PARAKH data both confirm, is that implementation quality at the school and classroom level varies enormously — and the variance is not random. The worst implementation is in the schools serving the most marginalised children.
Part Two: The Global Evidence on What Works
Teaching at the Right Level: The Most Robustly Evidenced Education Intervention in India
Pratham's Teaching at the Right Level (TaRL) methodology is, by any measure, one of the most rigorously validated education interventions ever developed. Since 2001, it has been evaluated through at least 11 randomised controlled trials in collaboration with J-PAL researchers at MIT — including Nobel laureates Abhijit Banerjee and Esther Duflo. The Global Education Evidence Advisory Panel has named TaRL one of just three "best buy" interventions for foundational learning globally.
The core insight is counterintuitive only in retrospect: instead of grouping children by age or grade and teaching the prescribed curriculum, TaRL assesses children by their actual learning level and groups them accordingly for targeted instruction. A Class 4 child reading at Class 1 level gets Class 1 instruction, with materials and activities calibrated to where she actually is — not where the system assumes she should be.
The results across 11 RCTs are consistent. In India, intensive learning camps of 10–20 days produced gains of up to 0.7 standard deviations — among the largest effect sizes ever recorded in rigorous education research. In Haryana and Uttar Pradesh, TaRL embedded within the government school system during regular school hours produced significant gains in the proportion of children achieving Grade 2 reading competency. Youth Impact, a rigorous implementer of TaRL, found the approach delivers 1.85 years of learning gains in a 30-hour programme. Among 27 studies of programmes to improve learning with cost data, J-PAL ranks TaRL among the most cost-effective.
There is one important caveat that the evidence is honest about: TaRL results are variable across contexts. A model that produced large gains in Uttar Pradesh produced no measurable impact in Assam — same organisation, same approach, same country. The evidence points to strong academic support for teachers, monitoring that functions as coaching rather than punishment, and genuine commitment from the government school system as the conditions that distinguish high-impact from zero-impact TaRL implementation. The approach is not a magic bullet. It is a validated methodology that requires the right conditions to work.
For Odisha, the relevant evidence is specifically that: NIPUN Bharat and FLN implementation that incorporates TaRL principles — ability grouping, level-appropriate materials, regular assessment — is on a stronger evidentiary foundation than any alternative approach. The question is not whether to do it, but how to do it well.
What this means for Odisha: NGOs supporting foundational learning in tribal blocks should use TaRL principles as their operating framework. This means: assessing children at the start of the programme using simple ASER-style tools; grouping by learning level for targeted daily practice; using oral and activity-based approaches that work in low-literacy environments; and tracking individual progress closely enough to catch and address the children who are still not moving. This is not different from what SCERT Odisha has designed for the government FLN programme. The question is whether it is being implemented with the quality and consistency that the evidence shows is necessary.
Mother-Tongue Education: The Non-Negotiable for Tribal Contexts
The research on mother-tongue-based multilingual education (MTB-MLE) is unambiguous and has been for two decades. Children learn better — faster, more durably, and with greater comprehension — when instruction begins in the language they already speak. For tribal children in Odisha who speak Gondi, Kharia, Santali, Kui, Bonda, or any of the dozens of other tribal languages spoken across the state, beginning formal schooling in Odia or Hindi is cognitively equivalent to learning to read in a foreign language. They are not just learning to read; they are simultaneously decoding sounds, words, and concepts in a language they do not yet understand.
The evidence on what happens when mother-tongue instruction is provided is consistent across contexts: stronger reading outcomes, lower dropout, higher comprehension at secondary level, and greater confidence. UNESCO has advocated for MTB-MLE for decades; the National Education Policy 2020 endorses mother-tongue instruction for the first five years of schooling.
Odisha has invested substantially in this. The SAMHATI programme and the development of FLN materials in tribal mother tongues represent genuine, evidence-based policy commitments. The state has tribal language materials in several major languages. The challenge is teacher capacity: a teacher who speaks Odia but not Gondi cannot deliver mother-tongue instruction in Gondi however well-designed the materials are. In blocks where teachers are transferred frequently and do not speak local tribal languages, the MTB-MLE policy exists on paper and not in classrooms.
What this means for Odisha: NGOs supporting tribal education have a specific and high-value role in the gap between MTB-MLE policy and practice. This includes: identifying and training community volunteers as language bridges between children and formal instruction; documenting where teacher–community language mismatches are occurring and supporting advocacy for local teacher recruitment; and building community libraries and learning materials in tribal languages that extend learning beyond school hours.
Community Volunteers and Extended Learning: The Evidence
One of the most consistent findings in education research is that learning does not have to happen only in school, and that community volunteers — with relatively light training — can produce meaningful gains in foundational literacy when given clear materials and ongoing support.
Pratham has demonstrated this across multiple volunteer-led TaRL programmes. Tamil Nadu's Illam Thedi Kalvi programme — which deploys community volunteers to extend learning beyond school hours — is an example at scale of what volunteer-supported learning can produce. Research on peer tutoring and structured community learning programmes consistently shows positive effects on foundational skill acquisition.
The relevance for Odisha's tribal communities is specific: in remote blocks where school functioning is inconsistent, where teacher attendance is low, and where the gap between government commitment and classroom reality is largest, community-based learning support is not a complement to the government system — it is sometimes the most functional learning that happens in a community. NGOs that build this infrastructure carefully — training village youth as learning facilitators, providing structured materials, and connecting to both community accountability and government monitoring — are producing impact in the spaces the formal system does not consistently reach.
What Does Not Work (And Is Still Being Done)
Infrastructure without pedagogy does not improve learning. The research on school construction, textbook provision, and even computer provision is consistently sobering: these inputs matter at the margin, but they do not produce learning gains without concurrent changes in how teaching happens. A smart classroom in which a teacher uses the same "chalk and talk" approach, teaching an age-grouped class the prescribed curriculum regardless of whether students understand it, produces the same outcomes as a classroom without a smartboard.
Teacher training that is one-shot and not followed by ongoing mentoring or coaching does not produce sustained change in classroom practice. The evidence on this is strong: a five-day teacher training workshop, without subsequent in-classroom support and regular follow-up, produces almost no measurable impact on student learning outcomes. What does produce change is sustained, school-based coaching where teachers receive regular feedback on specific classroom practices — from a mentor who has actually observed them teaching.
Government school rationalisation — the merging of small schools — has a documented negative impact on learning in remote areas, where merged schools may be several kilometres away, increasing the barrier to attendance particularly for younger children and girls. The CAG audit of Odisha's education department (2025) documented discrepancies between habitation mapping and actual school availability, raising concerns about the adequacy of access in some geographies even as enrolment numbers look good.
Part Three: Five Organisations Doing This Exceptionally Well
1. Pratham (India)
Pratham needs no extensive introduction in this context — it is, by a significant margin, the Indian education organisation with the most robustly validated approach to foundational learning. Founded in 1995 to improve education quality, it developed TaRL in the early 2000s and has since partnered with state governments across India to scale it, reaching more than 76 million students nationally. ASER — the Annual Status of Education Report — is Pratham's contribution to the evidence base: the first and still the most comprehensive national assessment of what children actually know, as distinct from what they are enrolled to learn.
Pratham's model for government partnership is instructive. They do not deliver education instead of the government — they work with government teachers, government school systems, and government monitoring structures to improve the quality of what is already happening. The shift in how teachers group children and what they do with each group is, in effect, a pedagogical transformation delivered through the existing system rather than parallel to it.
The transferable lesson for Odisha: The NIPUN Bharat FLN programme already incorporates TaRL principles. NGOs with Pratham's approach as their model should be working with government teachers and block-level officials — not setting up their own separate programmes. The value is in improving implementation quality within the government system, not in creating alternative institutions.
2. Agastya International Foundation (India)
Agastya works on hands-on, curiosity-based science and creativity education for government school students in rural India. Their model — mobile science labs, science fairs, and creative science instruction delivered in village settings — is not about FLN but about the question of what happens after foundational skills are in place: how do children develop the curiosity, critical thinking, and creative problem-solving that equip them for secondary education and beyond?
Agastya's evidence base shows consistent impacts on scientific curiosity, student engagement, and teacher confidence. In the Odisha context, their approach is particularly relevant to the secondary dropout challenge: children who find school engaging and who develop genuine curiosity are more likely to persist. The quality-of-secondary-schooling question — which is where tribal dropout is most acute — needs both foundational skill remediation and a more compelling vision of what school offers.
The transferable lesson for Odisha: Foundational learning and intellectual engagement are complementary, not sequential. NGOs working with tribal students in upper primary and secondary school need to be asking not just "can they read?" but "do they want to be there?" The latter question involves making school genuinely interesting — and that requires approaches like Agastya's, not just more intensive FLN drilling.
3. Vidyaposhak (Karnataka)
Vidyaposhak works in government schools in Karnataka on school community engagement — building partnerships between schools, parents, and village communities to improve school functioning and student learning. Their approach is relevant to Odisha because it addresses one of the structural weaknesses of government schools in tribal areas: the disconnect between the school as an institution and the community it serves.
Research on school community engagement consistently shows that when parents are meaningfully involved in monitoring school functioning — not just attending parent-teacher meetings, but having genuine voice in understanding and addressing why teachers are absent, why children are not learning, and what improvements the school needs — learning outcomes improve. In tribal Odisha, where community–school relationships are often distant or adversarial (because the school's language, curriculum, and culture are not the community's), this engagement is both harder and more important.
The transferable lesson for Odisha: School Management Committees exist in every government school, mandated by the Right to Education Act. In most tribal schools, they exist on paper and meet infrequently. NGOs that invest in making SMCs genuinely functional — helping community members understand their authority, track teacher attendance, monitor midday meal quality, and advocate for school improvements with district education offices — are producing the school accountability that learning outcomes depend on.
4. Teach For India (India)
Teach For India places high-quality graduates as full-time teachers in government and low-fee private schools in urban India. Their direct relevance to rural tribal Odisha is limited — they are primarily urban and require infrastructure that remote tribal blocks cannot provide.
But TFI's model offers a transferable concept: the idea that bringing highly motivated, trained, and supported people into classrooms — whether formally as teachers or as volunteers and community educators — can produce disproportionate learning gains. The key elements that make TFI teachers effective — intensive training, regular coaching, explicit learning goals for each child, and data-driven instruction — are the same elements that make any education programme effective.
The transferable lesson for Odisha: The NGO sector in tribal education can design fellowship-type programmes that place motivated, trained young people — ideally from local tribal communities — as community learning facilitators in partnership with government schools. This is not a substitute for properly functioning government schools; it is a bridge while those schools are still developing.
5. Eklavya (Madhya Pradesh and tribal India)
Eklavya has worked in tribal education in Madhya Pradesh and other states for over four decades. Their signature contribution is the Hoshangabad Science Teaching Programme — a model of hands-on, experiment-based science teaching for government school students that produced documented learning gains and influenced national science curriculum design.
More relevant to Odisha is Eklavya's approach to curriculum design for tribal contexts: developing materials that build on local knowledge, local ecology, and local cultural references rather than transplanting urban or non-tribal curriculum into tribal classrooms. Their research and practice on this — on what happens when curriculum is genuinely contextualised rather than superficially translated — is the most practically relevant body of work for NGOs developing educational materials for Odisha's tribal communities.
The transferable lesson for Odisha: Contextualisation is not just about language. It is about whether the examples, stories, and frames of reference in educational materials actually connect to children's lives. A maths problem about urban water tanks does not connect to a child in a forest village in Malkangiri. One about millet measurement or river fish counts does. The curriculum gap in tribal education is partly about this: materials that are linguistically translated but culturally and contextually alien still produce the cognitive load of learning in a foreign world.
Part Four: The Odisha Adaptation
What Transfers Directly
TaRL methodology transfers directly and is already embedded in NIPUN Bharat/FLN policy. The implementation question — how to ensure that teachers actually practice ability grouping and level-appropriate instruction, not just acknowledge it in training — is where NGO support is most valuable. The model of NGO as implementation quality support, rather than NGO as parallel education provider, is both more effective and more sustainable.
Community volunteer learning programmes — after-school, during school transitions, in the evenings — transfer directly in contexts where school functioning is inconsistent. They require trained and supervised facilitators, structured materials (ideally in local languages), and a genuine monitoring system to work. The evidence is clear that unsupported volunteer programmes, without materials and monitoring, produce little.
School Management Committee activation transfers directly. The legal framework exists. The skills to use it — understanding attendance records, tracking learning outcomes at school level, advocating with district education officials — can be built with relatively light but sustained NGO support.
What Requires Significant Adaptation
Teacher training and mentoring models designed for urban or non-tribal government schools require significant adaptation for tribal Odisha. The distances between schools make cluster-level mentoring difficult to sustain at quality. The language gap between mentor and teacher (who may speak Odia) and teacher and student (who may speak a tribal language) creates complexity that standard teacher training designs do not address. Effective teacher support in tribal blocks needs to be more intensive, more frequent, and more language-aware than standard government training designs.
Assessment tools — including the ASER oral reading assessments that form the basis of TaRL — require adaptation for tribal language contexts. An assessment tool in Odia does not tell you whether a Gondi-speaking child can read Gondi. The development of simple, reliable oral reading assessment tools in tribal languages is a genuine technical gap that NGOs with linguistic capacity can help fill.
What Must Be Built
A functioning pipeline of teachers from tribal communities is the most important medium-term education investment that Odisha does not yet have at adequate scale. The language mismatch between teachers and students in tribal blocks is fundamentally a teacher recruitment and training problem. Until there are substantially more teachers from the Bonda community teaching Bonda children, from the Kondh community teaching Kondh children, the MTB-MLE policy will remain aspirational. Building this pipeline — through pre-service teacher training incentives, scholarship support for tribal youth in teacher education, and strong local hiring preferences in tribal block postings — is a multi-year policy advocacy challenge for NGOs.
Longitudinal tracking of tribal students across school transitions — from primary to upper primary, from upper primary to secondary — is not happening systematically. The data exists in UDISE and district systems; the analysis that would identify which children, in which blocks, with which characteristics are most likely to drop out at secondary transition, and why, does not. NGOs with data capacity can build this analysis and use it to target preventive interventions — mentoring, scholarship support, school quality advocacy — at the transition points where dropout risk is highest.
Part Five: Government Scheme Mapping
NIPUN Bharat (National Initiative for Proficiency in Reading with Understanding and Numeracy): Central government's flagship FLN mission — targets universal achievement of Grade 3 literacy and numeracy competencies by 2026–27. Implemented through SCERT Odisha via Monthly Cluster Level Meetings for teachers, FLN materials, and DIKSHA platform. NGO role: supporting implementation quality at school level, community awareness, assessment facilitation.
SAMHATI (Odisha): Odisha's mother-tongue-based multilingual education framework for tribal children. Develops FLN materials in tribal languages and trains teachers. NGO role: complementing SAMHATI with community language support, tribal language literacy volunteers, and materials in languages not yet covered.
RTE Act 2009 (Right to Education): Mandates free and compulsory education from Class 1–8, prohibits automatic promotion to be abandoned (now modified under NEP 2020), requires School Management Committees. NGO role: SMC activation and capacity building, out-of-school children identification and mainstreaming.
EMRS — Eklavya Model Residential Schools: Central scheme providing free residential secondary education for tribal students in PVTG-concentrated areas. Odisha has 11 fully functional EMRS as of 2023. NGO role: community awareness, scholarship support, mentoring for EMRS students.
Ashram Schools and Sevashrams (Odisha): 705 Ashram Schools for secondary-level ST students and 501 Sevashrams for primary — government residential institutions integrating local languages and vocational elements. NGO role: quality monitoring, hostel condition advocacy, learning outcome tracking.
ANWESHA, AKANKHYA, KHUSI Schemes (Odisha): State schemes for scholarship support, girl child education incentives, and menstrual hygiene management linked to educational participation. NGO role: awareness, enrolment support, and monitoring in tribal blocks where scheme penetration is limited.
Mukhyamantri Medhabi Chatra Protsahan Yojana: Scholarship for meritorious tribal students. NGO role: awareness and application support in remote communities.
PM POSHAN (Midday Meal): Free midday meals in government schools — a critical retention and nutrition incentive. NGO role: community monitoring of meal quality and regularity, linking school kitchen management to Mission Shakti SHGs.
Part Six: Further Reading
Foundational Learning:
- Teaching at the Right Level — J-PAL Evidence Summary — Abdul Latif Jameel Poverty Action Lab: The definitive summary of TaRL evidence across 15 years of RCTs. Start here for anyone designing FLN programming.
- What Rigorous Evidence Reveals About TaRL — AfLEARN / DataFirst (2026): The most current and honest assessment of where TaRL evidence is strong and where it is context-dependent. Useful corrective to over-claiming.
- ASER 2023 — Annual Status of Education Report — Pratham: The foundational reference for understanding India's learning outcomes landscape. Read alongside Odisha-specific PARAKH data.
Odisha Context:
- Tracking Regional Disparities in Learning Outcomes — ORF (2025): Uses PARAKH 2024 data to map learning outcome gaps by state. Situates Odisha's performance in the national landscape.
- Education in Odisha — Grokipedia (2026): Comprehensive data compilation on enrolment, dropout, infrastructure, and ST/SC education programmes. Useful reference for programme designers.
- Building Inclusive Futures: Education Reform for Odisha's Vulnerable Communities — Tribal Research Institute / GoO (2025): Documents Odisha's own ST&SC education reform outcomes including the tribal female literacy improvement.
Mother-Tongue Education:
- Multilingual Education: From Policy to Practice — UNESCO: The foundational document on MTB-MLE evidence and implementation. Essential for anyone designing educational materials for tribal communities.
A Final Note: The Urgency Is Real
A child who leaves primary school without being able to read or do basic arithmetic is, statistically, set up for secondary dropout, low-skill employment, and the perpetuation of the poverty cycle into the next generation. The global evidence is clear that this problem is solvable — that targeted, level-appropriate instruction, delivered consistently, by a trained and supported facilitator, can produce rapid and durable learning gains even in children who have been left behind for years.
Odisha has the policy framework. NIPUN Bharat, SAMHATI, EMRS, and the residential school system represent genuine, evidence-informed commitments to tribal education. What they do not yet have, consistently, is the last-mile implementation quality that converts policy commitment into children who can read.
This is precisely where NGOs have their highest leverage. Not building parallel education systems — the government school is there, the teacher is (mostly) there, the curriculum and materials are being developed. But ensuring that in each school, in each tribal block, the FLN programme is happening with the quality and consistency the evidence shows is necessary. Monitoring teacher attendance. Activating SMCs. Training community volunteers. Building the language bridges that government policy designs for but cannot always deliver.
The scale of the learning challenge is not a reason for despair. It is a reason for specificity. We know what works. The work is implementing it — consistently, in the hardest-to-reach places, with the communities that have been failed the longest.
Written by the JaBaSu Trust knowledge team. Published April 2026. Corrections and additions welcome: knowledge@jabasu.org
Evidence Grade: A/B — Multiple RCTs (TaRL evidence is Grade A; broader learning outcomes evidence is Grade B). Last reviewed: April 2026
Related Knowledge Commons Content:
- Practice Note: Teaching at the Right Level — How to Implement TaRL in Tribal Schools
- Practice Note: School Management Committees — Making the Accountability Structure Work
- Practice Note: Mother-Tongue-Based Multilingual Education — From Policy to Classroom
- Org Spotlight: Pratham — Two Decades of Evidence-Based Education Reform
- Org Spotlight: Agastya International — Curiosity as a Retention Strategy
- Sector Primer: Child Welfare (Sector 01) — Education as child protection
- Sector Primer: Social Justice & Tribal Welfare (Sector 03) — EMRS and tribal education rights
Government Schemes Referenced:
- NIPUN Bharat / SCERT Odisha FLN Programme
- SAMHATI (Odisha MTB-MLE)
- Right to Education Act 2009
- Eklavya Model Residential Schools (EMRS)
- Ashram Schools and Sevashrams
- PM POSHAN (Midday Meal)
- ANWESHA, AKANKHYA, KHUSI Schemes
- Mukhyamantri Medhabi Chatra Protsahan Yojana