Quality of education is essential to the achievement of students. This is the central thesis of two PodCasts with Eric Hanushek (left) offered on EconTalk (I am catching up on the archive, forgive me). Hanushek has written extensively on this topic and I read a number of his papers for my Honours Economics thesis back in 2004 when I was trying to comment on the South African adoption of Outcomes Based Education and the problem of resources, school and teacher quality.

Roberts and Hanushek discuss everything from the PISA tests (which for some reason South Africa hasn't participated in), differences in test scores worldwide and the impacts of education on growth, to the impact of individual teacher quality on the achievement of individual learners. One of the fantastic points was that if kids in Mexico were brought up to the level of math and reading capability of US students, then Mexico should experience an increase in growth of approximately 2%. This is phenomenal. Which obviously makes me wonder, what would happen if we had a similar increase in the quality of education in South Africa?

First off, I don't believe that simply altering the system of instruction from one to another (such as with the OBE transition) is going to do it. Hanushek and others have shown that there are all kinds of things that affect educational outcomes (here have a look at the work of Esther Duflo, Chris Udry and Pranab Bardhan).

One of the problems that I often found with Hanushek's work is that he discusses US work in the reduction of class size for to decrease learner:educator ratios (which is PC for student:teacher ratios the phrase they use in the US, but SA uses L:E ratios as far as I recall). Anyway, his research indicated that there aren't significant gains from a reduction in class size (not T-S ratios, but that was also a topic elsewhere) from say 23 to 16 or 17. This is all well and good, but I don't see how this is going to be in any way informative in terms of the effects of reducing a class size of 50 to a class size of 30 or even 24 (Hanushek's high range above). I think that for this research, some valuable input has been provided by the paper by Angrist and Lavy on Maimonides Rule, which is a particularly important paper on the Economics of Education in the past decade. Anyway, they made a more important comment that a class size over 40 is detrimental to educational outcomes using a regression discontinuity design and instrumental variable estimation.

Combine the above with weird phenomena in South African education regressions and you begin to wonder about South African education and what kinds of inputs could be used. Moreover, there are continued disparities across income levels. Could SA run randomized evaluations in order to improve the data accessible to those trying to understand education policy? Yes! Could the department of education make data more accessible to researchers? YES! (My word have you ever tried to get hold of the School Register of Needs Data?). One of the things I wish we had is better data on teachers, schools and individual students (family background, detailed demographics, etc), then we could better measure the impact of teacher quality on student achievement and try to isolate what teacher quality is (it seems to be independent of experience and education level of the teacher by the way).

All of this has led me to believe that there is still a phenomenal amount that can be done in SA education. I still don't think that OBE is 'the answer'. I sincerely wish that there were more randomized evaluations and more research into possible other strategies for impact evaluation of programs. Maybe I just don't know about them, maybe they're out there and I haven't been told. I just wish that this kind of stuff was more readily available and promoted better for us to have a more rounded and scientific understanding of what we can do in SA in order to improve educational outcomes. This is one of the most crucial ways for us to attempt to correct for inequality (more on which in later posts soon).