Most primary school children in Odisha — particularly in tribal districts — sit in a Class 4 classroom but cannot read a Class 2 text. The formal system keeps promoting them regardless, so by Class 5 they are catastrophically behind, and by Class 8 they have likely dropped out. TaRL's response to this is radical in its simplicity: assess every child on what they can actually do right now (not what the curriculum says they should be able to do), group them with other children at the same actual level, and teach that group using activities calibrated to where they are. A child reading at letter-recognition level gets letter activities. A child reading paragraphs gets paragraph activities. As children progress, they move up. No child is left sitting through instruction so far above their level it is meaningless.
Component 1: The Assessment
What it is
TaRL begins with a simple oral reading assessment conducted one-on-one with every child. The ASER (Annual Status of Education Report) tool is the most widely used version: the assessor shows the child progressively harder reading tasks — letters, simple words, a short paragraph, a longer paragraph, a story — and notes where the child can no longer read fluently. A parallel numeracy assessment runs from number recognition through simple operations.
The whole assessment takes five to seven minutes per child. It is not a test with marks and a pass/fail. It is a diagnostic to understand where instruction should start.
Levels (standard TaRL framework)
For reading: Beginner (cannot identify letters) → Letter (identifies letters, cannot blend into words) → Word (reads simple words, cannot read connected text) → Paragraph (reads short paragraphs) → Story (reads fluently)
For maths: Beginner (cannot recognise single-digit numbers) → Number recognition → Subtraction → Division
Doing it in tribal Odisha
The standard ASER tool is in Odia. For children in tribal blocks who speak Gondi, Kharia, Santali, Kui, or another tribal language, an Odia-language reading assessment does not measure their reading ability — it measures their Odia language exposure. These are not the same thing.
Before you assess:
- Establish what language the children speak at home. Ask teachers, ask community members, ask the children directly.
- If assessment materials are available in the community's language (SCERT Odisha has developed some SAMHATI-aligned materials in selected tribal languages), use those for the baseline.
- If not, assess in Odia for your programme baseline — but document the language context prominently, and expect that children from non-Odia-speaking homes will cluster at lower reading levels than their actual cognitive ability warrants.
- Re-assess in the community language where materials exist, or through simple verbal comprehension tasks that do not require the child to read.
Practical tips:
- Train assessors to conduct the assessment in a friendly, relaxed manner. Many children in tribal blocks have had limited positive interactions with adults who are specifically paying attention to what they know. The assessment can be frightening if poorly handled.
- Assess in a quiet space away from classrooms — children being watched by peers will underperform.
- Do the assessment over two or three days if needed — tiring a child mid-assessment produces false results.
- Involve the school teacher as an observer, not as the assessor, initially. Once teachers trust the process, they can take over.
Component 2: Grouping
What it is
After assessment, children are grouped by their current learning level — not by their age, not by their grade, not by their caste or family background. A Class 3 child reading at Word level and a Class 5 child reading at Word level are in the same group for TaRL sessions.
This is the component that feels most counterintuitive to teachers and parents when they first encounter it. The system is built around grade progression. TaRL says: for the purpose of foundational learning, age-grade grouping is part of the problem.
How grouping works in practice
Groups are typically: Beginner group, Letter group, Word group, Paragraph group, Story group. In a school with 60 children across Classes 3–5, you might have 20 Beginners and Letters combined, 25 Words, and 15 Paragraphs and Stories combined. These ratios vary enormously by context — in some high-deprivation tribal blocks, 80% of children are at Beginner or Letter level.
TaRL sessions happen for a defined period each day — typically one to two hours — during which the school's regular timetable pauses and all children participate in their level-appropriate groups simultaneously. Outside these sessions, the school functions normally.
One important caveat: Groups are not permanent. As children progress — which, with good TaRL delivery, happens faster than you expect — they are reassessed and moved up. A child who moves from Letter to Word level within six weeks should feel that movement. It is one of TaRL's key motivational mechanisms. Movement is visible and celebrated.
Grouping in tribal Odisha
In small tribal schools with multi-grade classrooms (a teacher managing Classes 1–4 in one room, common in remote blocks), grouping is both easier and harder than in larger schools. Easier because the teacher is already managing multiple levels. Harder because there is only one adult.
Practical solution: Train two or three community volunteers — educated youth from the community, often available — to co-facilitate TaRL sessions. One adult manages the Beginner and Letter groups (which need the most active facilitation). The teacher manages the Word and Paragraph groups. The Story group often self-directs with minimal support after some practice.
Component 3: Teaching Learning Materials
What they are
TaRL materials are low-cost, locally appropriate, activity-based, and level-specific. They are not standard textbooks. They are:
- Picture cards with familiar scenes — a village market, a river, an agricultural scene — that prompt discussion, vocabulary building, and reading exercises
- Letter and syllable cards for letter-level groups to practise recognition and blending
- Word cards and simple text strips for word-level groups
- Short paragraphs and stories in local contexts for paragraph and story groups
- Number charts, bundles of sticks, play money for numeracy — concrete materials that make abstract concepts physical and manipulable
The principle behind the materials: children learn best when instruction connects to their own experience and knowledge. A maths problem about a railway timetable means nothing to a child who has never seen a railway. A maths problem about weighing millets for the weekly haat means something immediately.
Making materials for tribal contexts
SCERT Odisha has developed NIPUN Bharat FLN materials, including some in tribal languages through the SAMHATI programme. Use whatever exists. But for the specific needs of TaRL — level-appropriate, activity-based, using local cultural references — most of what exists in government systems is not quite right.
What you will likely need to make yourself:
- Picture cards using scenes from the specific communities you are working with — not generic rural India images, but recognisable local landscapes, crops, animals, activities
- Word lists derived from the words children actually use in daily life, not from Odia-language textbooks built around coastal or urban vocabulary
- Simple stories written in the community's context, ideally with inputs from community members and teachers who know the oral storytelling traditions
Cost: Materials for a classroom of 30 children typically cost between ₹500 and ₹1,500 if locally produced. They do not need to be professionally printed — handmade cards work fine, and there is evidence that children engage more with materials that look handmade and classroom-scale rather than polished and distant.
A practical starting point: Pratham publishes its core TaRL materials as public goods on their website and through the UNICEF FLN Resource Hub. Download and use these as the base. Contextualise the picture content and word lists for your specific tribal community.
Component 4: The TaRL Session
What a TaRL session looks like
A TaRL session is visibly different from a standard government school class. Desks and chairs are pushed aside. Children sit on the floor in their groups. The session is active, fast-paced, and frequently noisy in a purposeful way — children are talking, reading aloud, counting together, writing on the floor.
The structure of a session:
- Whole-group warm-up (5–10 minutes): All children together — a song, a game, a quick rhyme — that creates energy and signals the session has begun
- Level group activities (30–40 minutes): Each group works with their level-appropriate materials. The facilitator circulates — spending the most time with the Beginner and Letter groups, checking in on Word and Paragraph groups, monitoring Story group self-direction
- Individual practice (10 minutes): Each child practises independently. The facilitator moves through the room providing one-on-one support to children who are lagging or progressing
- Whole-group close (5 minutes): A shared activity that brings all groups back together — often another game or song — ending the session on a high note
What makes the session work
The facilitator never stands at the front and lectures. TaRL is not teacher-led instruction in the conventional sense. The facilitator is circulating, observing, encouraging, correcting, noticing.
Children teach each other. Children at the top of a level group are encouraged to help those just below them. This is not laziness on the facilitator's part — peer learning is one of TaRL's consistent sources of accelerated progress.
Mistakes are occasions for learning, not shame. In many government school classrooms, public mistakes are humiliating. TaRL sessions have to actively build a different culture. This takes time and consistent modelling by the facilitator.
The session is daily. Intermittent TaRL — once or twice a week — produces intermittent results. The evidence on TaRL consistently shows that intensity matters. Daily sessions, even for just 45 minutes, produce substantially better outcomes than equivalent time spread across a week.
Component 5: Mentoring and Monitoring
This is the component that most NGO-led TaRL programmes underinvest in, and it is the component most responsible for the difference between TaRL that works and TaRL that doesn't.
What the evidence says
J-PAL's twenty years of TaRL research are unambiguous: training teachers in TaRL and then leaving them to implement it does not produce learning gains. What produces learning gains is training teachers AND providing ongoing, on-site coaching from someone who has practised TaRL themselves for at least 15–20 days.
The key phrase is "on-site." A mentor who reviews data remotely and emails feedback is not a TaRL mentor. A mentor who visits classrooms, sits in on sessions, demonstrates better techniques, talks through data on individual children, and celebrates progress is a TaRL mentor.
Pratham's language for this is "leaders of practice" — people who have run TaRL themselves, understand what good implementation looks and feels like, and can coach others from direct experience.
What this means practically for NGOs
Your programme coordinator needs to be a practitioner, not a manager. If your programme coordinator has never facilitated a TaRL session themselves — has never sat on the floor with a group of Beginner-level children using letter cards — they cannot be an effective TaRL mentor. Build this into your programme design: your education coordinator spends the first month running TaRL sessions themselves before they begin mentoring teachers and volunteers.
Mentor visit frequency: In the first six weeks of a TaRL programme, mentors should visit each implementation site weekly. After six weeks, fortnightly. The visit has a structure:
- Observe a session (15–20 minutes minimum)
- Discuss what you observed with the facilitator — specific, descriptive feedback, not evaluative judgment
- Review assessment data together — which children are progressing, which are stuck
- Demonstrate a technique if needed (not lecture about it — demonstrate it)
- Set a specific focus for the next two weeks
Data for mentoring: At every mentor visit, bring the assessment data for every child in the programme. The conversation is not "are the children learning?" in the abstract — it is "Sunder is still at Letter level after six weeks; what have you tried with him? Let's look at what's happening."
Monitoring and review
Beyond individual mentor visits, TaRL programmes need a review cycle — typically monthly — where all facilitators in a programme area come together to:
- Share what is working and what is not
- Review aggregate data on learning level progression
- Problem-solve specific challenges
- Practice facilitation techniques together
These reviews are not reporting meetings. They are practice communities — the same model that produces better outcomes in teacher professional development everywhere.
Component 6: The Language Question — The Specific Adaptation for Tribal Odisha
This component is not named in the standard TaRL framework because TaRL was developed primarily in Hindi-medium and Odia-medium contexts where the language of instruction matches the community's language well enough. In tribal Odisha, it needs to be named and addressed explicitly.
The problem
A child who speaks Gondi at home and is being taught to read in Odia is doing two things simultaneously in a TaRL session: learning to read, and learning a second language. These are both possible, but conflating them in a single session makes both harder. The TaRL evidence base assumes that children have sufficient exposure to the medium of instruction to benefit from reading instruction in it. For many tribal children in Odisha, that assumption does not hold.
What to do
Step 1 — Establish the language baseline. Before beginning TaRL, document what language each child speaks at home. Do not assume. This information changes your implementation design significantly.
Step 2 — Use community volunteers as language bridges. A community volunteer who speaks both the tribal language and Odia can serve as a real-time bridge in TaRL sessions — explaining a word, a concept, or an activity in the tribal language when children are confused. This is not a substitute for mother-tongue instruction; it is a practical bridge that makes Odia-medium TaRL accessible to children who would otherwise participate with almost no comprehension.
Step 3 — Where SAMHATI materials exist, use them. SCERT Odisha has produced FLN materials in several tribal languages. For any communities where these materials exist, TaRL-style ability grouping in the tribal language for the early months, transitioning to Odia as children build confidence, is more effective than Odia-medium instruction from the start.
Step 4 — Advocate for local teacher recruitment. This goes beyond what a single NGO programme can solve, but it is the right advocacy goal: teachers who speak the community's language produce dramatically better outcomes than those who do not. Document the language mismatch in your programme data and use it to make the case to the Block Education Officer and District Project Coordinator.
What Success Looks Like — And When
The timeline
TaRL is not magic. In well-implemented programmes with daily sessions, good mentoring, and appropriate materials, you can expect to see:
- 6 weeks: Children at Letter level beginning to blend syllables into words; Beginner-level children reliably recognising most letters; overall classroom energy and engagement visibly higher
- 3 months: Measurable upward movement in learning levels for most children; some children completing two or three level transitions; Story-level children reading with confidence
- 6 months: Significant reduction in the proportion of children at Beginner and Letter levels; the most dramatic gains among children who had the furthest to travel
The evidence from Haryana and UP showed that at the start of a well-implemented programme, only 34% of Grade 3–5 students could read a simple paragraph or story. After one academic year, this rose to 53% in TaRL schools versus 48% in control schools. That gap — five percentage points at the aggregate level — understates the effect for individual children who moved from Letter to Story level within a year.
How to know if it is not working
If after six weeks you see: children still sitting in age-grade arrangement during TaRL sessions; facilitators standing at the front and explaining rather than circulating; no regular re-assessment of children's levels; mentor visits fewer than once per fortnight — the programme is not implementing TaRL. It is implementing a generic "classroom activity" with TaRL branding. Restart with the components.
Common Questions
"Our school has only one teacher. How can we do TaRL?" The single-teacher problem is real and common in remote tribal blocks. Solutions: trained community volunteers who co-facilitate (often local youth with Class 10 or 12 education); older Story-level children who facilitate Beginner and Letter groups under adult supervision; a "whole school" TaRL session timed to when both the teacher and the Anganwadi worker or ASHA can be present.
"The teacher doesn't want to change how they teach." This is the most common implementation challenge. Teachers who have spent years teaching in a certain way often experience TaRL as a criticism of what they have been doing. It is not. The approach that works: never frame TaRL as replacing what the teacher does — frame it as an additional session that the teacher runs, during which the regular curriculum pauses. Once teachers see children who were disengaged start engaging, resistance usually softens. But it takes time, and it takes a good mentor relationship.
"How do we work with the NIPUN Bharat / SCERT FLN programme?" TaRL is fully compatible with NIPUN Bharat goals — in fact, ability grouping and level-appropriate instruction are core to the NIPUN framework. Position your TaRL work as implementation support for NIPUN Bharat, not as a parallel programme. This makes government alignment much easier.
"What if parents object to their child being put in the 'lower' group?" Address this directly in community engagement before the programme starts. Explain clearly: this is not a permanent categorisation. It is a starting point. Children move up as they progress. Some parents whose children start at Beginner level have children reading stories within four months — and that transformation is more persuasive than any communication you can do upfront.
Resources
Free and immediately accessible:
- Pratham's TaRL Implementation Guide — available on the UNICEF FLN Resource Hub (unicef.org/flnhub)
- ASER assessment tools in Odia — available from ASER Centre (asercentre.org)
- SCERT Odisha SAMHATI materials — available from scert.odisha.gov.in
- TaRL Africa's classroom methodology documentation — available at teachingattherightlevel.org (highly detailed, highly practical)
Contact for direct support:
- Pratham Odisha office — they can provide technical assistance and mentor training for NGOs implementing TaRL in Odisha
- SCERT Odisha FLN division — for alignment with state NIPUN Bharat implementation
In Summary: The Six Non-Negotiables
- Assessment — every child, one-on-one, before grouping
- Grouping — by learning level, not grade or age, and regularly updated
- Materials — level-appropriate, activity-based, contextualised for the community
- Session design — daily, facilitated, activity-based, not lecture-based
- Mentoring — on-site, by practitioners, at minimum fortnightly
- Language adaptation — explicitly designed for the specific tribal language context, not assumed away
All six. Not five. All six.
Related Knowledge Commons content: Education Sector Primer (Sector 04) · Practice Note: School Management Committees — Making the Accountability Structure Work · Practice Note: Mother-Tongue-Based Multilingual Education — From Policy to Classroom
Evidence Grade: A — RCT-backed. This Practice Note draws primarily on eleven randomised controlled trials of TaRL implementation across India and Africa. Implementation guidance draws additionally on Pratham's operational documentation and field experience from TaRL implementation in comparable contexts. Last reviewed: April 2026.
Questions or corrections: knowledge@jabasu.org