Originally published in The Guardian as 'Admission of error', Tuesday January 21, 2003

As A-level results improve and competition for places at prestigious universities grows fiercer, admissions tutors are struggling to find ways to encourage students from deprived backgrounds. Are universities justified in demanding higher grades from independent school pupils? New evidence suggests they may be, says Bahram Bekhradnia

There is a clear and strong association between A-level scores and subsequent achievement at university, as John Thompson and I showed (Education Guardian October 15). This should have come as no surprise: unless A-levels are a complete waste of time, we must assume that they measure something, and that what they measure is not wholly ephemeral.

These results were seized upon by some as demonstrating that universities should rely solely on A-level scores to determine their admissions. Those findings, of course, did no such thing, as they were formulated at an aggregate level, and made no attempt to identify differences between students with similar A-level performance. In particular, they did not explore the way the school attended might influence A-level achievement.

Analysts at Higher Education Funding Council for England have done some further analysis, and have subjected the data to some sophisticated modelling. The results will be presented tomorrow at the Fair Enough conference in London. They have looked in particular to see if pupils who attended schools with relatively poor A-level results on average perform better at university than their peers from better performing schools, when all other things are equal. And, conversely whether pupils from relatively high performing schools perform worse than their peers from poor performing schools, again, all other things being equal.

This turned out to be more complicated than might be expected. It proved particularly tricky to try and separate the effects of school performance from school type, in particular whether or not the school was from the state or independent sectors. They found that there was no clear and uni-directional relationship between a school’s performance on average and the expected performance of all its pupils in higher education given their A-level grades. That is to say, it is not possible to make a general statement that a pupil attending a school whose pupils on average achieve poor A level grades, would perform as well at university as a student from a better performing school who had better A-level grades. The Hefce analysts do not go on to set out consequences for admissions tutors, but the implication must be that there is no basis here to make lower offers to pupils from such schools.

However, there was a sting in the tail – and a very interesting one. Although there was no general ‘school effect’, they found, as others have previously, that there was a clear "independent school effect". That is to say, they found that there was a clear and marked difference between the performance of state school pupils and independent school pupils when they got university. On average pupils from independent schools with a given score at A-level performed significantly less well at university than their state school peers. The difference was not large but it was significant -- on average equivalent to about 2 A-level grades.

What might the implications of this be for admissions policies? That depends on what admissions policies are intended to achieve -- something largely not articulated, though most admissions staff would say that they are looking for academic potential, and that when they make an offer to one applicant in preference to another they do so because they think that they are likely to get a better class of degree. Put another way, they are unlikely to offer a place to candidate A in preference to candidate B if they believe that candidate A will get a worse degree. Hitherto, the primary information on which they have based this judgment has been crude A-level scores, sometimes informed by interviews. Some universities have recently attempted to modify this with information about school performance characteristics, though the provisional results of the work carried out at Hefce do not appear to provide any general justification for this. However, what this research does show is that institutions may well be justified in making more demanding offers to candidate from independent schools, who will on average perform less well than their state school peers with equivalent A-level grades.

Potential academic achievement does not of course have to be the basis for admission to university. There are some universities in the United States that make offers to relatively poorly qualified candidates because they judge that they have the personal and social attributes to go on to be notable figures in their community or in society at large. Others may have a mission to serve their local communities above all else, and will have a policy of giving priority to local candidates – even if less well qualified than candidates from further afield. Others yet might value sporting prowess, and offer places to academically less well qualified candidates with sporting ability. I suspect that some of these policies might in this country fall foul of the Human Rights Act, but the important thing is that admissions polices and criteria need to be clearly articulated.

I have not explored why the relative performance of independently educated students should apparently decline in higher education. Reasons may include a higher quality of education and the degree of preparation for examinations they benefit from, both being effects which disappear at university, but both of which will serve to inflate A-level scores. The results of this research do not tell us whether these are indeed the reasons, though, do potentially still have significant implications for admissions policies. Without needing to engage in social engineering, there would appear to be good grounds here for the most demanding universities to differentiate their demands depending on whether pupils attended maintained or independent schools. But even this approach would be somewhat crude. What is urgently needed is a diagnostic tool to help admissions staff to distinguish the academic potential of applicants, often with identical A level results. The Sutton Trust have been pioneering the use of the Scholastic Aptitude Test (SAT) in this country. This may or may not prove successful, but certainly something more is needed. Otherwise we will continue to be forced to improvise around A level scores alone.