IQ tests not the answer

Making entry to Oxford and Cambridge fairer will take more than IQ tests, says Steven Rose

The view from Cambridge's King's College bridge of Gibb's Building, the chapel and Clare College is one of the most sublime architectural harmonies I know. And yet, in my more radical fantasies, I have, if reluctantly, come to the conclusion that if we are ever to achieve a more egalitarian society in Britain, it may have to go. For however liberal the intentions of the universities of Cambridge and Oxford are, they remain, at the end of this millennium, the educational access points to privilege.

Better funded, rich in tradition and self-confident, with the right contacts in the more esteemed professions readily available, how could it be otherwise? To graduate from either is still to obtain a head start over those from most other English universities. Coming, as I do, from a north London Jewish background and a direct grant school, I know how much I too have benefited from this educational silver spoon.

Of course, both institutions are aware of this, and for decades have been embarrassed not by the fact, but rather that entry into these gateways to privilege was disproportionately from private as opposed to state schools. Widening access is one way of co-opting bright working-class youngsters, yet it has not proved easy. Decades ago I remember a fellow of one college, involved in selection, explaining to me how he bent over backwards to accept state school applicants, but they simply were not up to the quality of those from the private schools.

So how should one respond to the news that Oxford is now contemplating using "high-powered IQ tests" to spot potential in state school students?

I suspect that, knowing my scepticism about what such tests measure, The THES expected I would produce a root-and-branch denunciation. Well, up to a point. It is worth remembering that when the first IQ tests were produced, early this century, they were intended for precisely the purpose for which Oxford proposes to use them - to provide measures of a schoolchild's performance that were independent of a schoolteacher's assessment, and that would enable the child's performance to be compared with that of others of the same age. This would help identify those for whom additional teaching might be required.

In fact IQ measures correlate reasonably well with teachers' assessments of a pupil's ability and performance, and with exam results. This is scarcely surprising, because this is precisely what they are intended to do. The fact that IQ scores are distributed in the famous "normal" or bell-shaped curve is also unsurprising, because test items are chosen in order to achieve this result and to eliminate average gender differences in performance. Just as examinations are supposed to.

It is perfectly possible to set exams that virtually everyone passes, or that everyone fails. But as anyone who has sat on an exam board knows, this is not their intention. The British system demands that we distribute the exam scores with about 10 per cent firsts, 10 per cent fails and the rest in between, and if this is not the way the marking works out there is embarrassment all round and post-hoc adjustment to bring the results more into line.

The point is that such a distribution is a response to social expectation, not a biological fact of life, and in this sense IQ tests are no more and no less measures of some absolute capacity than are exam results. The chequered history of their use, tied to racism in the US and the class system in the UK, and in both with eugenic undertones, is enough to make anyone sceptical.

Throughout the century the claims have been, first, that the tests measure some real fixed "quality" in a person - the so-called "Spearman's g", or "crystallised intelligence"; and second that this quality is largely dependent on one's genes. These claims ignore the fact that the abstract numerology of IQ scores has rather little to do with real measures of life success or creativity. And further, that in the richly intricate jungle of brain processes that modern neuroscience is studying, there is no space for a reified lump of genetically fixed "intelligence".

I welcome any effort to expand the entry base into Oxford, but let us not assume that IQ selection tests are going to solve the problem of educational privilege for universities, or provide a rational biological base for dividing the population into sheep and goats, any more than Cyril Burt's use of IQ tests back in the 11-plus days did for secondary education.

Steven Rose is director of the Brain and Behaviour Research Group at the Open University.