Skip to content
The UK's only independent think tank devoted to higher education.

Above and beyond predictions – No exams presents an opportunity for innovation in contextual admissions

  • 7 May 2020
  • By Clare Marchant

This blog has been kindly contributed by Clare Marchant, Chief Executive of UCAS.

With A level results day now firmly circled in the sector’s diary for 13 August and guidance issued about how students will be awarded their grades, minds in admissions offices across the country will begin to look ahead in earnest to late summer – and the busy job of confirming applicants’ places dependent on upcoming results.

But now, possibly more than ever before, fairness for students must be ingrained into every possible aspect of the admissions processes. There’s a need to focus efforts on making sure that over a decade of work in widening access and participation doesn’t fall by the wayside. These often lesser supported students need more guidance and reassuring messages than ever as they journey through the next few months, spending more time out of formal education than any incoming undergraduate cohort in history.   

The Schwartz principles of 2004 still broadly ring true today. If you flip them around, you’d be hard pressed to find someone who thinks admissions should be shrouded in mystery, or we should use assessment methods that are unreliable. This has, this week, been reiterated by the UUK new proposed sector agreement on fair admissions practices. The enactment, and clear communication to applicants and their advisers, of fair admissions remains paramount, and with appropriate contextualised admissions policies in action, the applicant can be considered in the round, with their key background characteristics taken into account.    

There’s an argument that, for this summer’s applicants, the application of contextual admissions, underpinned by those principles, requires a still greater, albeit nuanced focus. So, as strange as it might sound, there are aspects of this cycle that are going to be very familiar to admissions teams. The ball isn’t completely in their court of course – context can only be considered if they know about it. A key piece of advice to give to applicants in any year is to stay connected with their universities, with admissions teams working to make sure that they are looking at an applicant’s full story and not just grades on a piece of paper. This is as true today as it would have ever been.

But what about those grades? Across the UK last year, nearly 13,000 18 year old UCAS applicants were under predicted in their A levels. That’s around 1 in 12.

These students, in more conventional times, would seize the moment when taking their exams and excel beyond their teacher’s predictions. It’s imperative that we as a sector do not overlook them. While some of our new survey work shows applicants are still fixed on their university selections and are keen to do what it takes to get to their dream university, some feel that they have lost their chance to shine by not sitting an exam.

These are the students that universities and colleges should now have at the forefront of their mind as the admissions cycle works through to its conclusion. These are the students who could’ve previously shone in a traditional exam setting. These are the students with a clear potential to flourish, but only if the sector plays its part. 

Within that 13,000, there are many stories to tell. Some may come as little surprise, such as applications for medicine and higher tariff universities being the least likely to be under predicted, but there are some compelling analysis we can glean from the rapid research conducted by UCAS.

Of those applicants who went on to achieve A*A*A last year, almost 39% of POLAR4 quintile 1 (the most disadvantaged) applicants were underpredicted, compared to just 27% of POLAR4 quintile 5, and this pattern holds true across achieved grade profiles of CCC and above. 

Young men are slightly more likely to be underpredicted than women (8.2% vs 7.7%), as are students from grammar schools or sixth form colleges (almost 10% of applicants), while students from non-selective state schools are the least likely to be under predicted (7%).

While unconditional offers can’t be ignored, any student with a confirmed place (regardless if they accepted an unconditional offer, or had their place confirmed when A level results are published) will again be able to release themselves into Clearing as last year to exercise their choice again – and so could move to a new university, college or course. 

Almost 6% of applicants with an unconditional offer at their firm choice are underpredicted, compared to more than 8% of those with a conditional offer. Any of these students could decide to be among the expected 50,000 that could use self-release this year. Applications to Art & Design and Mass Communications courses are most likely (just over 10%) to be overpredicted, likely linked to them being the subjects where unconditional offers are most prominent

There are some limitations with these numbers. We’re just looking at UK 18 year olds with 3 predicted and achieved A levels from 2019, and applicants at the top and tail of predicted grade profiles aren’t especially useful for analysis. There are also differences in proportions of applicants being underpredicted in each group, which could be down to differences in grade distributions as in the case of POLAR4 quintiles. 

While this year’s admissions cycle will forever be known as different, it also gives us a chance to build on 15 years of living and breathing fair admissions principles and apply them to a new, unforeseen scenario – especially as it was clear that the January deadline snapshot was painting a picture that 2020 was set to see the most material improvement in WP measures over the last 5 years.

An opportunity for innovation in contextual admissions, albeit not without challenge.  This could be the catalyst for a rebirth of fair admissions with positive lasting effects for years to come. 


  1. Alex says:

    If the baseline for this data is three A level grades, and many students from POLAR quintile 1 cohorts study BTEC courses, how accurate is the reflection of the landscape presented here ? With over 100,000 students a year now applying with BTEC qualifications, surely it would make sense to include them in the stats?

  2. Mark Corver says:

    Defining under-prediction as what proportion of those who got, say, “A*A*A” in exams but were predicted less can cause confusion in my experience. When you have two groups with differing levels of average attainment – like POLAR – you will get apparent ‘under prediction’ of the lower achieving group similar to that described. This is simply a result of the relative profiles of the attainment distribution and the natural randomness of grades. It doesn’t say anything about whether predicted grades are ‘fair’ for people from different backgrounds. But there is a risk that is what people conclude.

    A clearer way to assess fairness is to use the distribution of exam grades relative to predicted, not the other way around. We did urge UCAS six months ago to consider a simple matrix of predicted and achieved A level points by background for this purpose. It would be immensely helpful to both universities and schools if these data were published. Doubly so if the BTEC data were included.

Leave a Reply

Your email address will not be published. Required fields are marked *