Newsletter 67 SA Education System - A statistical analysis

Luis Crouch and Thabo Mabongoane published a statistical research paper in 2001 “No Magic Bullets, Just Tracer Bullets” asking;

  1. “Are resources the key to increasing the education system's performance, as measured by, say, matriculation pass rates?
  2. If one "controls," statistically, for resource availability, is the poor performance of many schools perfectly understandable?
  3. Or is there too much emphasis on trying to increase and redistribute resources, and too little emphasis on the wise management of those resources?
  4. Or can either resources or good management really do much to improve results, given the widespread social inequalities, poverty, unemployment, community and familial problems and crime?
  5. Is it right to expect that the education system, as such, will have much effect on learning, when there are such daunting social-environmental factors impacting on schools?
  6. It would seem critical to begin trying to develop rigorous answers to such questions.”

The researchers – making use of statistical models – issued the follow caution;

                “In spite of these problems the results are strong enough that their implications cannot be         ignored—they pass the economists' inter-ocular test: they hit one right between the eyes. Yet,    because it is only one study, and it has limitations, its conclusions should not have too much impact on policy until other studies, by other researchers, begin to confirm our findings, and until       it is possible to include all (or, say, 95%) of schools in the study. If other studies deny our findings,           then we have to go back to square one.”

Here are a summary of their findings

  1. Poverty is an important factor that cannot be subsumed under the notion of "resources." Even controlling for resources, poor children, or schools in poverty-stricken areas, tend to perform much worse than others: schools in very poor areas tend to have matriculation pass scores some 20 points lower than schools in richer areas, even if one statistically makes resources "equal." It is thus wise to be increasingly modest about how much one single sector— education—can achieve until the worst aspects of poverty are dealt with via economic growth, redistribution, targeted anti-poverty programmes, and programmes aimed at increasing social capital.
  2. Related to poverty but somewhat independent of it, and independent of resources, is the fact that a school's being in a township, or just being ex-DET, also appears to decrease matriculation pass rates by about 20-30 points. Again, even with more or less equal resources, and controlling for poverty, ex-DET schools appear to underperform. This suggests that managerial and culture of learning and teaching issues are of extraordinary importance: these factors seem to matter more than resources—certainly more than the "bricks and mortar" and easily quantifiable resources.
  3. When it comes to resources, the learner-educator ratio seems to matter much less than the quality of the educators. The qualifications of educators (as measured by the average REQV at the school) seemed, by far, a more important factor than any ratio, or any other cost-related resource factor. An increase in the education of educators equivalent to one REQV (roughly one year) appears to be associated with an increase in pass rates of about 16 or so points. The physical conditions of the school (as assessed in the SRN) did not appear an important factor, if one accounts for all the other ones we have mentioned, though these conditions probably do play some role.  Interestingly, we found little evidence that years of experience matter much, thus calling into question all the criticism of the re-deployment and rationalisation process as resulting in the loss of the most experienced teachers. Thus, the study begins to suggest that a process of vastly improved educator upgrade, support, and productivity supervision, may well be far more important than focusing on L:E ratios and quantitative equality. Other resources— such as computers and the availability of well-stocked media centres—appear of some importance. We doubt that computers and well-stocked media centres are the only important physical resources, and assume that they simply proxy the fact that kinds of resources that are strongly associated with immediate cognitive delivery (books, stationery, etc.) do matter.
  4. Finally, we found that, as expected, even after taking into account all our measures of poverty, resources, and so forth, we have that some 30% of the performance of schools remains "unexplained." We take it that this refers to managerial factors. Thus the importance of an emphasis on management. South Africa has done much too little on this score so far, and what little it is doing seems half-formed. In terms of management strategy (which our policy conclusions above suggest is important), we can conclude that it is possible, even now, to develop quantitative evidence, school-by-school, on what a reasonable level of expectation for each school should be (at least in the more advanced provinces—in the others it might take a few years). This can be used in management in various ways. Schools that are performing below expectation, even after controlling for the fact that they are in a poor area or have few resources, can be targeted for managerial rather than resource assistance. These will often be schools that are not really in extremely bad shape, from the point of view of quantitative resources. Schools that are performing reasonably well managerially, given their resources, but if they are in a poor area, can be favoured in terms of resources, because these would be the schools where, since management seems to be rather good and yet the area is poor, then improved resources are likely to have a more liberating effect, and are likely to be well-used, than in schools that are well-off or poor but are badly managed. We would feel confident in making the claim that good analytical and EMIS methodology for developing these targeted management interventions, while not perfect, can be applied in South Africa right now, at least in the more developed provinces. It will, however, require the development of more analytical skills in the staff, or more staff with such skills.”


If your school is “underperforming” remember

  • Poverty cannot be ignored
  • To succeed you MUST deal with broader socio-economic inequality first
  • Township and exDET schools are more likely to underperform
  • Teacher qualification is the most important resources
  • Teacher experience does not account for a significant resource
  • Teacher-pupil ratio account for less than teacher qualification
  • Management capacity accounts for more than 30% of resources.

Add Your Comment 

* Name:   
* Email:    
* Comment:    
  Please calculate the following and enter the answer below: 1 x 4 + 5 = ?
Please leave this box blank.