Admissions tutors should not wrestle with contextual data alone

Source: Miles Cole

While this is the time of year when Oxbridge applicants nervously await the outcome of their interviews, we should also spare a thought for the plight of the poor admissions tutor.

The annual agony of the admissions process, with interviews held right up until Christmas week, brings many difficult decisions, not least the fraught question of whether and how to use contextual data.

The issue is never far from the headlines and, this autumn, press attention turned to a report by the advice body Supporting Professionalism in Admissions (Contextualised Admissions: Examining the Evidence) which found that 37 per cent of universities were using contextual data and 57 per cent planned to do so. The Daily Telegraph’s spin on this was that leading universities were “secretly admitting poor students with lower entry grades”, leading to “renewed concerns that [they] are trying to ‘socially engineer’ the admissions process”.

Personally, I am surprised that only 37 per cent of universities use contextual data. Why would you not take into account relevant factors such as the performance of an applicant’s school, the average university attendance rate in their neighbourhood and the relative affluence of their area when deciding whether to offer a place or, in the almost unique case of Oxbridge, calling a candidate for interview?

Some admissions tutors argue that interviews do the job of contextual data by allowing them to mine for “rough diamonds”. But others believe that interviewing only helps to select privileged candidates (not all of them from private schools) with networks that provide opportunities to practise. Either way, constraints on time and space limit the number of interviews that can take place. Oxbridge application numbers are kept down by self-selection based on teachers’ advice, but many other leading universities have 20 applicants per place for subjects such as law and medicine. For them, interviewing every candidate is not realistic.

Some contextual information is helpfully flagged on the application forms we pore over (although we also need to be sure that we are comparing students who have taken comparable examinations; given the bewildering array of exam boards, and exam types, this is quite a challenge). But how much account should we take of it? Should we, for instance, make a lower-than-standard offer to pupils from disadvantaged backgrounds?

Some universities insist on the same offer being made to all, but tutors in institutions that do not do this have a difficult decision to make. Who could fail to be impressed by a candidate from such a background whose grades were just a little worse than those of someone with an elite education? But mathematical subject tutors in particular tend to feel that students simply will not cope with the early stages of the most demanding courses if they have not achieved top grades at school. I sometimes think that the only solution is a billionaire donor who would fund four-year degrees at top universities for these students who would benefit from a foundation year before they merge with other students on the standard three-year degree.

Either way, it seems unrealistic for each admissions tutor to have sole responsibility for factoring in contextual information, not least because of the time pressure they are under. Far better would be for Ucas to develop a standard mathematically derived formula to guide us as to how much leeway we have in making lower offers to candidates from deprived backgrounds. This would be a considerable task for Ucas to undertake but it would do the sector a great service.

The issues around contextual data are muddied by several facts about our current, suboptimal admissions system. For instance, admissions tutors have to make decisions based on predicted rather than actual grades. But in October, OCR, one of England’s biggest exam boards, revealed that only 48 per cent of predicted grades are accurate. It is also possible that the accuracy varies with subject, school type and so on. For now, why not make offers based on uniform-mark-scale marks and scale these according to contextual information?

All students should be assessed on exactly the same information. Many foreign applicants’ Ucas forms do not provide the requested information about the type of school they attended and I am increasingly concerned that, while we are committed to outreach and widening participation among home students, we may unwittingly be admitting only a social elite from abroad.

Equally unfairly, many foreign students do not declare age-16 exam results (indeed, their home countries often have no GCSE equivalent). In contrast, for home students we rely heavily on GCSE results, often stipulating a minimum number of A*s. This means that we are not giving our own young people much opportunity to right past wrongs and to refocus their goals post-16. My best advice to students who are performing much better than their GCSEs would indicate (or who think that they will do better than their teachers predict) is to defer application until after their A-level results. Why the rush? Their generation will probably work until they are 70, so waiting a year seems a small price to pay to attend the best university they can.

As for admissions tutors, we tend to self-flagellate over these issues, but fellow sufferers might care to read Daniel Golden’s brilliant book, The Price of Admission, which provides a damning indictment of the way elite US universities favour, among others, applicants who are white, excel at sport or are the children of alumni, donors or celebrities. We in the UK have room for improvement, for sure, but we are not that bad.