What 4,000 reading assessments told us

Since 2018 we have run an ASER-style oral reading assessment twice a year on a sample of Classes 3 to 5 in every school on our register. It takes about four minutes per child: can she read letters, words, a Class-1 paragraph, a Class-2 story. We now have something over four thousand of these. The first finding was the one we expected, only worse. In our baseline year, 69% of the Class 5 children we assessed could not read a Class-2 level paragraph. Not "read slowly" — could not read. The state average is bad; our schools, which we chose precisely because they are the hardest, were worse. The second finding was not expected at all. When we disaggregated the 2018 results by school, one school in Khammam had a cluster of children who could read letters fluently and then fell apart at word level, which is not how reading failure normally looks. Our volunteer that day, a medical student, asked one of them to read the same passage held twelve inches from her face. She read it perfectly. We screened that school and found 23 children out of 148 with uncorrected refractive error. In the district-wide screening we ran the following year across eleven schools, the figure was 14%. Spectacles cost us ₹340 a pair at the volumes we buy. That year, spectacles moved more children up a reading level than books did, and it was not close. This is why we run health camps, and it is why the Nutrition & Health programme sits inside an education charity rather than beside one. We did not set out to be in the eyesight business. The data put us there. The current cohort figure, across adopted schools, is that 58% of children can read a Class-2 paragraph by Class 5, up from 31% in 2021. We publish the full assessment methodology and the raw district-level figures in the annual report, including the schools where the number went down.