THE LETTER FROM F Coetzer (“Standard deviations”, 3 December 2010) refers. Coetzer responds to the explanation of the admissions policy at the University of Cape Town (in a letter from Gerda Kruger, 4 November) by attempting to reinforce columnist Stephen Mulholland’s previous assertion that by admitting some students with lower matric marks than others, the University of Cape Town (UTC) will produce groups of “mediocre” students and “brilliant” students. He argues that, statistically – on the basis that there are fewer students in the interval between 91% and 100% (the interval for accepting white applicants) than between 74% and 100% (the interval for accepting black applicants). On those grounds he argues that because there’s a normal distribution of results in both groups there must be more mediocre students in the black group admitted to UCT.
He’s wrong. In fact, because of the extreme educational disadvantage faced by most black pupils their normal distribution is significantly shifted to the left, such that there are in fact fewer black pupils scoring above 74% than there are white pupils scoring above 91%. That’s precisely the root of the problem – and the evidence of disadvantage and the reason UCT applies different admissions cut-offs.
We believe talent is randomly distributed in all race groups, in all schools and geographical areas, whether rural or urban. We believe the performance distribution within any school system or given set of environmental determinants is a consequence of the distribution of talent, plus motivation, hard work and, no doubt, some opportunities. But, between school systems and environmental settings, different performances are much more influenced by those differing systems. Therefore, to attract the brilliant in all settings one should simply select the top 5% of students, regardless of their actual marks. That’s essentially what UCT’s system is doing. Not surprisingly, we find the top 5% from one school system will have marks that are much higher than the top 5% from a different system.
Unfortunately, we also find the performance of the top 5% in some systems is too poor; they just haven’t mastered enough during their 12 years of schooling and the university is unable to make good that deficit, even though we recognise these applicants have the raw ability and motivation to succeed given the right support. So we can’t admit them.
But it’s just this statistical thinking (as described by Coetzer) that results in the admission of students from different socio-economic communities and different school systems on the basis of a different set of marks. The proof we’re getting it right is that once they’re given the support UCT can offer – in residence, with tutors and mentors, and perhaps an extra bridging year of study – these students do as well as their peers with the high matric marks.