School quality and educational outcomes in South Africa

Data are drawn from the 1993 South African Living Standards Survey (SALSS), supplemented by test-scores obtained from specially designed math and comprehension tests that were administered to a fraction of the respondents in conjunction with the SALSS. The study also uses administrative data on pupil-teacher ratios at the level of magisterial district, and merges these into the SALSS and test-score data. Even at the district level, where data are aggregated across schools, there are wide disparities in pupil-teacher ratios, not only between s (19 pupils per teacher) and blacks (41 pupils per teacher), but also within the black population. The richness of the SALSS data and the study"s merging procedure allow examination of a wide range of inputs and outputs. The study finds that family background variables exert a powerful effect on educational outcomes for children, but also finds a strong and statistically significant effect for pupil-teacher ratios. These effects are not detectable for, and are much stronger for blacks when pupil-teacher ratios are high than when they are low. Low pupil-teacher ratios have a positive effect on test scores, school attendance, years of completed schooling at any given age, and parents" expenditures on their children"s education, so that private educational expenditures complement public educational expenditures, suggesting that parents will share the costs of educational improvements with the state. These results stand in sharp contrast to the often-expressed view that public expenditures have little effect on the educational attainment of children. (Author abstract)