This article is more than 3 years old

Admitting mistakes: should the government take control of university admissions?

A new report proposes a radical overhaul of university admissions. Debbie McVitty unpicks the policy bluster and the legitimate challenge.
This article is more than 3 years old

Debbie is Editor of Wonkhe

A new report from specialist independent think tank EDSK takes as its premise the notion that for university admissions to be fair, transparent and equitable, universities must sacrifice some autonomy over admissions.

The frame: university autonomy

University autonomy over admissions was, of course, reaffirmed in the Higher Education and Research Act 2017, which prevents the Secretary of State for Education issuing instructions on the criteria for admissions, or their application.

At several points in the report, the authors Tom Richmond and Andrew Bailey accuse universities of “hiding behind” the Act to “shield themselves” from government intervention on admissions. The basis for this claim is universities’ lukewarm response to Damian Hinds when, as Secretary of State, he wrote to universities urging an end to the practice of conditional unconditional offers – a dubious action on legal grounds, as Smita Jamdar pointed out on Wonkhe.

It’s rather unfair to accuse universities of cowering behind a law that was, after all, designed for their protection. But on the wider admissions debate, universities have demonstrably engaged – most specifically through undertaking a sector-led review of admissions mirroring the Office for Students (OfS) government-mandated review.

The principle of autonomy over admissions is currently being tested in a much more thorough way, with the outcome of the OfS consultation on the integrity and stability of the English sector expected imminently. And, once established as an acceptable point of intervention, it may be hard to roll back, even when the immediate crisis and associated risk to the student interest has passed.

Richmond and Bailey certainly see the OfS consultation as the basis for a future, much more regulated and nationally consistent system – why, they ask, rhetorically, would you not want a stable sector and protected students indefinitely?

For much of the sector, this will come over as fighting talk. But the reality is that autonomy is a spectrum, not an absolute state. And part of thinking through what an admissions system that can work across devolved nations, that respects student choice and maximises fairness – however you define that – will involve universities agreeing to sacrifice some autonomy, working within the UCAS admissions system being one notable way universities already do this.

So, in that sense, the idea that there could be some new ways for universities to collaborate or harmonise approaches is hardly shocking. And when you look beyond the proposal that the whole thing should be handed to OfS to regulate – the suggestions of how this might be done are certainly worthy of consideration.

The rationale

The three arguments the authors have with the current system are: the unreliability of predicted grades, the documented rise in unconditional offers and the failures of current practice to address inequality and disadvantage.

Predicted grades aren’t always the most accurate guide to ultimate exam performance; this is well established. But as Mark Corver has pointed out, predicted grades and exam grades are both trying to capture some third quality – potential for future success. Neither is precise and both bake in some level of educational disadvantage. l

When David Kernohan crunched the numbers for the 2019 admissions cycle he found that the rate of over-prediction was reasonably consistent, which led him to suggest that over-prediction is priced into the system – so even if it looks a bit skewed, it all comes out in the wash.

Ben Jordan, head of strategy at UCAS, points out that predicted grades will still be needed, even if they form no part of the admissions system – applicants and their advisors will need some sense of likely performance to calibrate university searches. But, he adds, that doesn’t mean they are a necessary feature in university admissions processes or that there aren’t other ways of expressing them to improve accuracy or better support academic judgements.

The arguments on unconditional offers are well-rehearsed – my own view is open unconditional offers should be encouraged, but I recognise I’m a bit of an outlier on that position. I do find the argument that some applicants should get lower offers in recognition of their disadvantaged circumstances – but absolutely not unconditional offers – a bit circular.

In the area of tackling educational disadvantage, it’s hard to argue with Richmond and Bailey. They make the point that qualitative elements like entrance tests, interviews, and personal statements are vulnerable to bias, and offer opportunities for well-resourced applicants to access additional support to make their case.

Contextual admissions, the authors argue, are approached inconsistently across the sector, with little clarity over the link between different forms of disadvantage and offer-making. You can see why different subjects in the one institution might have a different approach to contextual offer-making – due to the nature of the subject, particular access and participation goals in that subject, and so on – but it’s hardly straightforward for candidates. Add to that the potential for extra form-filling and hoops to jump through and it’s an obvious area for improvement.

The recommendations

Given that Richmond and Bailey appear to have limited confidence in the appetite of universities to update their own processes and achieve consensus on a consistent approach, it’s not surprising that top of the list of recommendations is that the Office for Students should be given responsibility for the regulation of admissions.

Universities would be required to subscribe to UCAS as a condition of registration, with UCAS acting as a designated body for admissions – a proposal that will raise a few eyebrows over in Cheltenham. “One of the benefits that UCAS brings to the sector is our independence,” says Ben Jordan. “We’re independent from government, and UK wide, operating in the interests of students – there is real value in this, so being a designated body may compromise that.”

The next set of recommendations focuses on published entry criteria – it’s suggested that every course publish a standard entry qualification, along with a maximum number that can be recruited to that course while maintaining quality. The standard entry qualification isn’t a terrible idea – in Scotland, the authors point out, moves are afoot for universities to judge what the absolute minimum qualification requirements are for success in any given course, and publish these. The range of entry qualifications and accreditation of prior learning would complicate matters, but that’s what the tariff system was designed for.

It does raise the question of whether it would be reasonable to publish no minimum threshold – the Open University would certainly have to be able to. In some cases, such as access to foundation years, the minimum threshold can be as low as having attempted Level 3 – this is typically targeted at mature students.

The maximum numbers idea is almost a throwaway proposition, but it’s an intriguing one. The ways that are available to universities to account for higher than expected recruitment in a given subject area are often unpalatable – hiring short-term teaching staff, increasing class sizes and so on. You could keep the consequences light – make recruitment beyond your published limit a reportable event, and ask universities to demonstrate how they’re maintaining quality, rather than taking punitive action for a relatively minor and low-risk infringement.

But on the other side of the argument, there’s certainly a risk, as with all student numbers caps, that the less advantaged get squeezed out. And the question of “what is a course” is a hardy perennial – each course is rarely a distinct entity with a fixed capacity.

Further to the standard qualification offer, the report proposes a nationally consistent approach to contextual admissions, in which various forms of disadvantage would be assigned points and those points would correspond to reductions of grades in offers. A national approach seems like overkill, but there would certainly be merit in universities agreeing to adopt a consistent approach to contextual admissions and publishing how these affect offer-making either at subject or institution level.

It would also make sense to find consensus on which factors are included in assessing for contextual admissions: care leavers is obvious, free school meals is also a good contender, arguably refugee or asylum seeker status should be added. Deprivation is a bit more complicated if you’re looking at individuals – coming from a deprived area isn’t always a great indicator. In Scotland the Scottish Index of Multiple Deprivation is the agreed standard – so something like the UCAS MEM could be helpful.

Now we come to the fun bit – allocation of places. Richmond and Bailey propose that predicted grades, personal statements, interviews and the rest of it be excised from the process entirely. Instead, students would rank up to ten courses in order of preference and would automatically be allocated to their highest preference course possible when they got their results. Where courses were oversubscribed, allocations would be made by lottery. And where students changed their mind or failed to be allocated to any of their courses, they would enter Clearing as usual.

The idea certainly meets the test of fairness and transparency. And, in a potentially positive way, it separates out the processes of universities and applicants getting to know each other from the process of selection – meaning that universities’ engagement with potential applicants would be solely focused on matching courses and students rather than on judging the candidates. It strips out the whole dance of offer-making from the system entirely. And, honestly, even at the selective end, if a student is academically qualified, how much does it really matter if they also have Grade 8 cello?

Sos Eltis, a fellow at Brasenose College, Oxford, was recently quoted as saying that applicants tended to put too much emphasis on being well-rounded: “You can be as flat as a pancake so long as you are good at the subject.” Though Oxbridge tutors would almost certainly point out that while A level performance is a useful starting point for selection at that level, they’re looking for something extra and beyond competence in exams.

Though it would be a radically different approach, I suspect that universities might be somewhat more interested in entertaining it than politicians. It would be politically very brave indeed to propose that the middle classes be stripped of many of their prospects for gaining an edge in the system – though a lottery system would still not address the educational inequities that mean that middle-class applicants are more likely to have better level 3 performance in the first place. But even so, it’s likely to make a lot of applicants and parents nervous, and would probably require some detailed polling and development before it could be implemented.

The report concludes with a section on all the areas the authors haven’t quite managed to think through – this is very much a report about school leavers in England. Mature students, international students, those with exceptional circumstances affecting exam performance, and so on, are deferred for another day.

The premise that admissions is static simply isn’t true – Ben Jordan points to adjustment, Clearing Plus, the UCAS Hub, and UCAS’s own ongoing work on predicted grades (which will support the wider reviews of admissions) as evidence that the sector continues to update its thinking on admissions.

In this respect, the EDSK report fails to give universities the benefit of the doubt. This is a mistake – even if only because it views as settled the position of the sector and thus defers to government and regulation as the answer, when much progress could be made on a voluntary basis. Beyond the question of university autonomy, some of the ideas proposed are definitely more radical than you might expect from a sector-led admissions review. So on that level, it’s a useful contribution to the debate.

It’s also a salutary reminder that public trust and confidence in university admissions is a core feature of a civilised and equitable society. While it’s no doubt frustrating to have universities’ intentions second-guessed, the public must be able to trust that access to university is a function of your academic merit and potential, not your social background, or worse, your social connections.

As Richmond and Bailey put it:

It is deeply concerning how wealth and privilege continue to unduly influence who gets accepted onto university degrees, particularly at the most prestigious institutions. This inevitably results in an overwhelming sense of unfairness as well as risking a catastrophic loss of trust – not just in the admissions process, but in the education system as a whole.

8 responses to “Admitting mistakes: should the government take control of university admissions?

  1. This is a great follow up to Richmond’s work on grade improvement, where he saw a vast series of examining bodies setting standard exams to determine the number of firsts that could be awarded. Now he wants to set up a vast bureaucracy to capture entrance standards, predicated on the mobile middle classes wanting to maximise the ‘prestige’ of their university place.

    There are parallels with the Irish system in what he proposes – but they have key advantages over the English: it’s much smaller system with a broader school curriculum.

  2. Has any research ever been conducted into a ‘net gain’ model? Not 100% sure if that’s the correct term but what it essentially would mean is developing a system to see which students outperform their school environment? Eg a student who has achieved an above average grade at a ‘low’ performing school would be seen as achieving a greater result than someone with the same grades but at a ‘high’ performing school?
    High and low performing schools would have to be a reflection of access to HE and other significant factors. There would probably have to be an element of personal net gain based on some of the factors as mentioned in the article included as well.
    Not sure how this would work for students who have been out of education for a while but not sure that would nesesarily a new factor.

  3. A link to the report itself might be helpful:
    https://www.edsk.org/wp-content/uploads/2020/06/EDSK-Admitting-mistakes.pdf

    It appears the authors have done little original research or thinking. They favour PQA but are unable to name a single country that has a genuine post-PQA system and better access outcomes than the UK. They discuss inaccurate grade preductions but do not consider how this relates to marking accuracy and which grades are in fact most affected. They favour merging UCAS with OfS but do not stop to question why the UCAS monopoly should be retained at all and whether better alternatives exist. The section on unconditional offers essentially regurgitates higher education media contributions from other people (THES, broadsheets, and yes even WONKHE). There is scant consideration of the causes of the rise in conditional UOs and the treatment of the symptoms (yet more central control) lacks imagination. So for example, revoking offers for poorer than predicted grades or awarding module credit for good grades are not even discussed as remedies notwithstanding the extensive US experience of their use.

    Think tanks are supposed to think from outside the box.

  4. You say that this will be more popular with universities than with government, but I have my doubts.

    We have had years of increasingly transactional and “marketised” HE and across that period Recruitment and Admissions divisions have experienced huge growth in both number and power within their universities.

    If there’s one thing these reforms would do for universities, they would save them a huge amount of staff costs in exactly those divisions. You could comfortably lose 75% of the staff in year 1: several million per year for many universities, even before you look at indirect savings in reduced IT, buildings, HR etc.

    The question would be: are those divisions powerful enough to successfully petition for their own survival or would Finance, marking them down as a huge cost centre they can live without, successfully argue for them to be scrapped?

  5. @Michael: yes, there is a lot of research on this. An example that sets out results very clearly is https://dera.ioe.ac.uk/19811/1/HEFCE2014_03.pdf by HEFCE. School factors included type of school/college (e.g. independent vs state, FE college vs school etc) and average performance in terms of A-level grades (average points per entry). It didn’t directly look at the effect of schools’ HE entry rates but it did look at POLAR on the pupil level.

    The results are very interesting and I advise you to read the whole thing if you’re interested, but in summary:

    * A-level grades are an extremely strong predictor of degree outcomes, to a far greater extent than any other factor they investigated

    * After taking A-level grades into account, the effect of school performance is only significant if you look at the top 20% vs the rest – there is no significant difference between the bottom 4 quintiles. Students who attended a top 20% school perform about as well as students from other schools whose A-level results were one grade lower (i.e. a student from a top school with BBB is as likely to get a 2:1 as a student from a more average school who achieved BBC).

    * The gap between independent vs state schools is also equivalent to about one A-level grade for higher-achieving students but bigger for lower-achieving students (e.g. independent BCC is equivalent to CCD at a state school).

    * POLAR adds little information for most of the grade range. After you take A-level grades into account there is basically no difference in degree outcomes for POLAR Q1 vs Q5 students with A-level grades between ABB and DDD. But students from POLAR Q1 areas with very high or very low A-level grades tend to do worse than students with the same grades from POLAR Q5 areas.

  6. There is already widespread avoidance of UCAS mainstream processes – Mark Corver points to RPAs as being “the fastest growing route” representing 8% of all entries https://wonkhe.com/blogs/the-thickening-fog-of-the-ucas-rpas/; Music conservatoires have long had their own parallel system that although it sits alongside the mainstream system now and is administered by UCAS, it could have been entirely independent; and Scottish FE colleges simply don’t use UCAS for their large full time HNC/HND cohorts because it’s a waste of applicant money. No doubt there are other “niche” areas where UCAS is simply avoided for new full time undergraduates (without anyone even knowing, yet alone sanctions being taken), and of course it doesn’t apply to the admittedly diminishing until now, but still large in absolute terms, part time or professional body HE awards – which just may experience a lift in demand in the environment of mass unemployment shortly about to envelop us.

    The idea that Scottish universities or population would accept a takeover by OfS on behalf of the UK government, responsible solely for English domiciled students, is such a nonsense that it’s scarcely worth considering. We would just see a mass exodus back to the system of individual application.

    It’s ironic that the only thing at present that the current free market right wing conservative government might consider nationalising is … University admissions! There might be more than a few people who would point out the hypocrisy of this in Tory philosophy, and instead argue that the bankrupt private care home sector or the airline or bus industries just might merit a higher priority for state intervention in the current crisis …

  7. Notwithstanding the good work that the team at UCAS do to engage with students and student representatives, it’s surely a stretch for UCAS to claim that because they are independent of government or regulators they therefore are “operating in the interests of students”. A majority of the board are Vice-Chancellors. UCAS’s primary accountability is to universities, and sometimes it shows.

  8. The part of this I find the most problematic is the end of personal statements and interviews. I understand that they introduce bias into the equation, but without the opportunity for universities and applicants to get to know each other, I fear there will be more drop-outs. The personal statement and interview process is vital to assess that students understand the subject they are applying for and specifics of the course they are applying for. Otherwise, we get students who assume that our BA in X Subject is exactly the same in terms of content as a different BA in X Subject at a different institution. Or worse, they simply assume content based on their own understanding of the subject area. Our website is clear what our programme involves, but they either don’t read it, or it doesn’t override their preconceptions, or they really need to talk to someone about it to find out if it’s right for them. We can’t force them to come to an open day before they apply, so interviews are the only way to ensure they talk to someone. Admitting these students just because they’ve put us as their first choice will lead to them dropping out. We know this because its happened before when an admissions tutor hasn’t interviewed or been really strict about personal statements. We admit them, they come, they’re not happy, they drop out. Not good for anyone involved.

    Maybe a post-offer mandatory interview is an option – they aren’t allowed to make the decision to accept until they have had a conversation with a member of teaching staff from the programme, and then it’s up to that member of staff to make sure that student fully understands what they are getting into.

Leave a Reply