Oxford University's admissions system is arbitrary, unfair, unjust and ought to be radically overhauled, Professor Alan Ryan, the Warden of New College, Oxford, said yesterday.
He said applicants should take psychometric tests to identify those likely to flourish, and colleges should lose the right to decide who should be admitted or rejected. One problem was that tutors differed so much in what they were looking for that a candidate's chances of being accepted varied considerably from one college to another, Professor Ryan said.
"This matters less in subjects such as maths, where - not very surprisingly - the Oxford intake looks exactly like the intake of other top universities, and matters more in subjects like English and history, where it looks very different," he said. "As a matter of justice, applicants in a given subject ought to face the same standards across the university and it is obvious that in many subjects they don't."
Another "obvious injustice" was that the numbers applying to any given college in any given discipline differed dramatically from year to year. "It seems unfair that the amount of competition a candidate faces should vary so arbitrarily with the accident of which college she or he has put down as a first choice," he said. The system of placing rejected candidates into a pool for other colleges to pick up was merely palliative. The real problem was that there was "too little predictability about what candidates must do to get admitted". Professor Ryan argued in the Oxford Magazine that one part of the solution was to introduce psychometric testing to help differentiate between candidates who would nearly all have a minimum of three grade As at A-level and A*s and As at GCSE.
The tests, which are widely used by employers, would help to identify such qualities as energy, ambition and self-reliance, which marked the difference between those who flourished at Oxford and those who did not. Responsibility for admitting students to the university should then be handed over to a central panel of "admitters", who would "tick off" GCSEs, school reports, test results and other available written work.
Once admitted, students would choose which college they wanted to go to and tutors would decide which of the admitted students they wished to teach. Professor Ryan said: "It would be extremely unlikely that any student would be thought admissible by a panel and unacceptable by every single college." He said the result would be a vastly better system than that operated by Harvard or Princeton, where he was previously Professor of politics.
A psychometric test that identifies "fluid intelligence" and is not affected by pupils' home background or by how well they were taught has been tried successfully on 150 Oxford applicants.
The test was devised by Dr Jane Mellanby, of the university's department of experimental psychology, to counter the charge that Oxford was biased against state school pupils. It was taken voluntarily by the sixth-formers under examination conditions with the assurance that the results would not contribute to the success or otherwise of their applications.
First, they had to answer five questions on a short passage about the Green revolution which contained any "fairly simple" graph. Condensed from an article in Nature, it was chosen because it "related to the sort of problem of which intelligent sixth-formers in any subject would be aware and it did not favour science over arts graduates". The first question asked the candidates to describe the graph, while the four others were designed to test their ability to think beyond the data they had been given.
Next, they tackled a questionnaire containing 30 statements that they had to say were "very unlike me" or "very like me" on a scale of 1 to 10. This was designed to identify their "learning style" - deep learning, which maximises understanding and is related to long-term professional success, or surface learning, which is a form of rote learning and is not associated with long-term success.
Dr Mellanby said the most important finding was that the scores on the Nature part of the test were similar for candidates from state and independent schools even though their GCSE results differed significantly. "Our test therefore seems to rely less on the knowledge base that independent schools are more practised at teaching than state schools, and therefore it may be more culture-free," she said.
The second finding was that pupils who did well on the Nature questions also had high scores for "deep learning". That suggested the two parts of the test were an accurate indicator of the "fluid intelligence" - or novel problem-solving ability - that the university was seeking to identify. Dr Mellanby said the findings of the trial, which was funded by Peter Lampl's Sutton Trust, justified a larger study and pointed to a new and far more accurate method by which to select undergraduates.