Summary
Key findings
- About six years after the programme, learners whose Grade 1–3 teachers received on-site reading coaching scored 0.12 SD higher than the control group on a composite of six Grade 8–9 subjects (significant at the 10% level).
- Learners whose early-grade teachers received coaching outperformed the control group in almost all secondary-school subjects examined; the differences were statistically significant only in Mathematics (about 3.0 percentage points higher) and Economic Management Sciences (about 2.1 points).
- There was no significant effect on English First Additional Language or Setswana Home Language marks, suggesting the earlier Setswana literacy gains (0.17 SD in Grade 7) may have faded by Grades 8 and 9.
- Coached learners were 7.2 percentage points more likely to reach Grade 9 on time, a 17% increase over the control group, of whom 43% were on track.
Methodology & context
- The Early Grade Reading Study began in 2015 as a randomised controlled trial with 4,538 Grade 1 learners in 230 no-fee (Quintile 1–3) schools in two districts of North West Province, aimed at improving early-grade reading in Setswana; this paper compares the coaching arm (50 schools: lesson plans, graded readers and specialist on-site coaches) with 80 control schools.
- Learners were linked to 2023 School Based Assessment records (Data Driven Districts system): 2,794 were matched and in high school, and 1,533 with marks in all six subjects form the main sample; effects are estimated by OLS with baseline reading scores, other controls and strata fixed effects, with standard errors clustered by primary school.
- Results are intent-to-treat effects for learners who could be matched (22% of control and 19% of treatment learners were not matched, a difference that was not statistically significant); a Grade 11 follow-up using direct learning assessments is expected to give stronger evidence.
Summary drafted with AI from the AFEP working paper and policy brief, and checked against them. See the full study for details.
Abstract
Programme models intended to improve learning at scale have increased in low-and middle-income countries. Although some of these programmes have led to gains in learning outcomes, very little research has been conducted on the extent to which the effects of these scalable models and programmes persist, especially beyond initial improvements in early grades. Using evidence from South Africa’s Early Grade Reading Study, a randomised control trial with three treatment arms implemented from 2015-2017, we assess whether gains in foundational reading skills achieved in the second or third year of schooling translate into better learning outcomes in secondary/high school education. By means of an estimation equation employed in a previous evaluation of the study, we evaluate administrative data to show a sustained impact of the intervention, with treated learners outperforming their control group peers by 0.12 standard deviation in overall high-school subject performance in Grades 8 and 9. The results from this study support the argument that investment in foundational skills during early grades can lead to long-term improvements in learner outcomes, particularly through the coaching intervention approach implemented in the Early Grade Reading Study.