#1216 2001 · Pratham (Rukmini Banerji) · Education / primary literacy and numeracy
Pratham stopped teaching by grade and started teaching by what a child actually knew
问题
Grade-five classrooms held children spanning six years of reading ability, so grade lessons taught most of them nothing
背景
India's push for universal primary enrollment succeeded at getting nearly every child into a classroom seat, but the curriculum still moved in lockstep by grade and age. Independent assessments later confirmed what teachers already suspected: roughly half of rural grade-five students couldn't read a simple grade-two-level paragraph, with actual reading levels inside a single classroom spanning five or six grades' worth of ability.
The standard response to low learning outcomes — more textbooks, more teacher training on the existing grade curriculum, smaller class sizes — still assumed the grade was the right unit to teach to. Every one of those fixes kept aiming instruction at a level most of the room couldn't use, leaving the furthest-behind children with lessons they had no chance of following and the furthest-ahead children bored by material they'd already mastered.
换别人会怎么做
Every mainstream lever assumed the grade level itself was the correct unit of instruction and tried to make it work better: distribute more textbooks, retrain teachers more thoroughly on the same grade curriculum, or shrink class sizes — all of which left instruction still targeted at a level most students in the room either couldn't reach or had already surpassed.
他们看到了什么
'Grade five' was an administrative label, not a measure of what a child could do. Pratham measured real ability with a five-minute test, then taught to that measured level instead of the assigned one.
那一手
Starting with pilots in Mumbai from 2001, Pratham built a simple oral assessment any volunteer or teacher could administer in minutes, then used it to sort children in grades three to five into groups by their actually demonstrated reading and math level — not by their grade or age. Each group received instruction, activities and materials matched to that level for a dedicated block of the school day, with children re-tested and re-grouped regularly as they progressed, so the classroom's teaching unit became a child's measured skill rather than the calendar year they were born in.
为什么管用
A quick, repeatable assessment turns an invisible, assumed-uniform classroom into a visible spread of measured skill levels, which can then be regrouped and re-taught to directly. Because the assessment is cheap and fast, children can be re-tested and re-grouped often, so instruction keeps tracking a child's actual, moving skill level rather than freezing them in whatever group they started in — closing the gap between what's taught and what a child is ready to learn far faster than a fixed grade-level curriculum ever could.
值了多少
Evaluations found learning gains up to 0.7 standard deviations; the approach now reaches over 50 million children across India and Africa.
什么时候会失灵
The approach needs a school system flexible enough to physically regroup children across their assigned classes for at least part of the day, which rigid, single-teacher-per-classroom structures or strict grade-promotion rules can block. It also depends on frequent, low-stakes reassessment being logistically and culturally acceptable — systems where testing is expensive, infrequent, or treated as a single high-stakes gate rather than a routine diagnostic tool lose the continuous feedback the model relies on to keep matching instruction to ability.
后来呢
Teaching at the Right Level is now cited by J-PAL and the World Bank as one of the most cost-effective, evidence-backed interventions in development economics, adopted in partnership with more than half of India's state governments, and scaled internationally through TaRL Africa, launched in 2019, which reached five million children across sixteen sub-Saharan African countries within five years.
资料来源
- [1]Tailoring instruction to students' learning levels to increase learningAbdul Latif Jameel Poverty Action Lab (J-PAL), 2021povertyactionlab.org
- [2]Mainstreaming an Effective Intervention: Evidence from Randomized Evaluations of 'Teaching at the Right Level' in IndiaNational Bureau of Economic Research, 2016nber.org