Higher Education: A way out of the university maze: Setting a list of criteria should be merely a first step in selecting successful candidates, writes Michael Collins

If it is thought that the present system of 'clearing' is rough justice, then at least for university applicants who perform as they expect, there is certainty. It serves few if such a system is replaced by one devised to serve those who administer rather than those who consume.

In 'Last straw for university lottery' (4 July), Tony Higgins, chief executive of the Universities and Colleges Admissions Service (Ucas), described two possible models. Ostensibly, both make sense; after all, why leave so much to uncertainty when a computer can allocate candidates according to their preferences, examination results and other criteria set by institutions? Or else institutions award places in a carefully timetabled sequential operation after examination results are known, albeit at the price of a later start to the academic year?

But whom do these proposals serve? To quote Mr Higgins, describing the virtues of the first model: 'This type of system . . . guarantees that courses meet their targets precisely, provided sufficient qualified applicants are out there.'

Notice that there is no mention of individual choice here, either on the part of applicants or of universities; after all, the Higher Education Funding Council for England will have decreed how many students may be admitted to which universities and in which subjects. But this is not the end of the story. The algorithm proposed in the Ucas consultative document is fatally flawed.

The algorithm first scans all first choices made by applicants to see if places on each course can be filled, for which it is necessary only that an institution's criteria are met. If more applicants achieve the required 'score' than there are places, the best are chosen. Then second choices are considered, and so on. This would seem eminently fair, but is it?

Consider an applicant who, seeking the most demanding course and so qualified, just misses out (perhaps by the Ucas lottery). In the second round, he or she may be excluded again if that course has had its places filled in the first round by first-choice applicants.

The same applicant will also be excluded by the second proposed system. Thus our unsuccessful applicant has been penalised solely for having the ambition to seek the challenge of the most demanding course.

By first constructing a simple model for the Ucas proposal and seeing how to adapt it, one can show that no single automated system can be foolproof. The target must then be to minimise the possibility of an injustice.

One approach that would require neither a delay in the start of the academic year nor a dependence on conditional offers, is a variant of the second option, but closer in practice to the present system since some version of 'clearing' will always be necessary when the initial choices of applicants, even when qualified, may not match the places available.

Candidates submit their applications after Christmas, listing, say, six or eight courses. At this stage, schools will have a better idea of their progress than with the present timing.

Any university is free to make an unconditional offer of two grade Es, but such an offer lapses unless accepted before some specified date, at which time all remaining applicants specify an order of preference.

Once A-level results are known, universities process the remaining applications, knowing applicants' order of preference (unlike at present). First, each university eliminates all applicants whom it knows it will not accept, either because they are not 'qualified' or because all places on the chosen course can be filled by first- choice applicants, and applicants' orders of preference are revised.

Only then is the acceptance process carried out and the process repeated with universities able to consider second- choice applicants and so on. This way, each applicant receives, at most, one offer. After this first round, universities publish their remaining vacancies, applicants again submit a list of preferences, and the whole process is repeated: the point is that the lottery aspect of 'clearing' has been removed.

This procedure would be fair to both applicants and institutions. It requires more work, but it is no longer dependent on determining 'scoring systems' in advance.

There is far more to selecting candidates for admission to a university than setting a specific list of criteria. Candidates are individuals, and no mechanised system can account for those qualities that are unquantifiable.

Where do we insert those crucial judgements of the ability and motivation to make a success of future study at university as opposed to past record, or of those qualities that can only be assessed in relation to other applicants?

Universities must be allowed to continue to care for their students, and the efforts to select to the best of our abilities are well rewarded, provided that we feel more than a computed affinity with those whom we teach.

The author is tutor for admissions at University College, Oxford, where he is a Fellow and Praelector in Pure Mathematics.

(Photograph omitted)