Is university admission an academic decision?

One topic which exercises many universities is admissions: not only for the obvious reason of recruiting enough students to meet targets, but also for the question about who should be in charge.

Across UK higher education, the underlying culture is that it is an academic decision: suitability to study for a programme should be determined by the academics who teach the programme.  This doesn’t mean that actual academics always take decisions, however: many universities have agreed that specific decisions can be taken by professional service staff, as long as they fall within parameters agreed with admissions tutors. So, if a student gets more than so-many tariff points, or better than such-and-such A level grades, they can be offered a place without reference to a tutor.

David Willetts (in his very interesting book, A University Education) reminds us that the UK is odd in this regard. In the US, admissions decisions are not typically made by faculty tutors, or not even in consultation with faculty tutors. Decisions can be based, for example, upon familial donations; upon siblings having attended; or on residency within a particular state. (Before you get too shocked, I recommend that you have a read of Willetts’ book: there’s more too it than nepotism and a disregard for academic standards.)

The difference can be understood, I think, in relation to a very good underlying principle, which is that academic decisions can only be made by academics in the discipline concerned. This is at the heart of academic freedom. Ask yourself a question: what is the academic decision which is at the heart of university admission?  Is it about who socially gets to do higher education? That doesn’t feel academic to me. Is it about whether a person has the necessary prerequisite knowledge? (For instance, do you need A-level maths to take the first-year modules on the programme?) That sounds much more academic, and is at the heart of the differences in the UK. In the UK specialism takes place at the start of university education; in the US students enrol, study a wide variety of modules for a couple of years, and then choose their specialism. And they take an extra year (at least) to study, so there’s time for this breadth.

It's that picture again! 

I don’t think its controversial to say that there are US universities operating this approach which are at least as good as UK universities. The UK systems generates good graduates a year sooner than the US system, but that isn’t because we’re cleverer: its because the system is structured to produce graduates after three years. As part of this, it is necessary to have early specialisation, and this means that admissions decision have to consider specific subject knowledge and readiness for study.

Now I am going to say something slightly controversial. These tests are more about the resources devoted to pre-university education and upbringing rather than any intrinsic academic merit. We know that a private school education boosts a person’s chances of getting good A-level grades and hence a place at a ‘better’ university. We also know that, in aggregate, for students with the same A-level grades, those educated at state schools will do better overall than those educated at private school (see, for instance, this HEFCE research). This means, I think, that private school – with better resources, smaller classes, and concomitant greater parental support for learning – has a better short-term impact.  But when learning resources and chances are evened out at university, the impact dissipates.

The point is that university entry based on A levels is about readiness to study. Background knowledge, confidence and social capital are what matters, because this enables a person to graduate in three years.

On this telling, university admissions should really be understood as a business decision. Remove some of the selective elements, and you won’t get the three-year throughput upon which the UK higher education system is built. (The development of foundation years to enable wider entry to selective universities supports this point: only by an extra year can pre-university educational differences be resolved.) University admission is only an academic decision because we set the system up to make it so. More time at university would enable foundation level study to become a norm. And at that point entry decisions would not be about pre-requisite knowledge, and entry barriers would come down.

And this is my challenge to the Office for Students, and to the UK government’s review of higher education. If you’re serious about removing social barriers to higher education participation, what are you going to do to enable longer degree programmes, to take the apparently academic decision out of the admissions loop?

Dreaming spires

Oxford University reports today (2 September 2016) that its 2016 entry has the highest proportion (59.2%) of state-school educated students for four decades. See the report from the BBC.

An Oxford College and two buses

No doubt this is good news, but it seems to me there’s a long way to go. Let’s have a bit of a further poke at this.

(Disclaimer: I went to state school and didn’t get a place at Oxford. But I’m trying my best not to have a chip on my shoulder about it. Going to LSE instead (again, a pretty elite institution) was the best thing that could have happened, and I really wouldn’t have thrived amongst the dreaming spires!)

Firstly, how what are the relevant populations? The most recent Department for Education data (nb this is for England only) give the following numbers of 16-18 year olds entered for at least one A level; I’ve added the proportion of the whole which this represents:

Students entered for at least one A level or Applied single/double award A level
State	153,034	59%
Independent	33,999	13%
FE College	74,468	28%
Total	261,501	100%

You’ll see that there are three categories of institution – state and independent schools, and FE Colleges. I haven’t been able to source the actual numbers from Oxford’s data, so don’t know the detail, but my assumption is that they mean that 40.8% of the 2016 intake come from the independent sector – that is, the 59.2% figure is the sum of all state-educated entrants. Why do I assume this? Because if they had entrants from FE colleges as well then it would be a stronger story that they’re becoming more diverse, and they’d have said so. So, 40.8% from independent schools it is.

This means that the 13% from the independent sector become over 40% of the total; and the almost 87% from the state sector become not quite 60%.

Oxford admitted just over 2600 UK-domiciled new undergraduates in 2014 (again, the most recently available data.) Assuming that this number stays about the same, then we can see how many independent and state-school educated students they will have, and compare it with what a proportionate distribution would look like:

	Percentage		Numbers
	Actual	National proportion	Actual	National proportion
State	59.2%	87.0%	1,539	2,262
Independent	40.8%	13.0%	1,061	338
Total	100.0%	100.0%	2600	2600

So the actual number of independently-educated students attending Oxford this autumn will be over three times the proportionate share of independent schools amongst those taking A levels - 1061 compared to 338. On a simple reading, you’re three times more likely to go to Oxford then you should be, if you go to an independent school.

What conclusions do I draw from this?

Firstly, there’s a manifest unfairness here. It is good that Oxford is doing something about it; and very good that that something is beginning to show results.  Three cheers for Oxford. And keep it up – you’ve a long way to go!

But the underlying unfairness is one which Oxford cannot be blamed for, and cannot address.  Independent schools are characterised mostly by their fee paying status. This enables them to have better facilities than state schools, and smaller class sizes. This in turn means that they are better at getting good A levels, and helping their students’ gain confidence. As long as this gap exists, between what state schools can do, and what independent schools can do, then we’ll see a difference in outcomes. And the difference in outcomes that we currently see is based primarily on family wealth.

I’m not trying to blame individuals for this, or make them feel bad. But as a society it seems that we’re selling ourselves short if we don’t enable all people to maximise their chances in life, and that for me means making state education better. A lot better. This is how we make society more equal, and probably better and more prosperous for all. 

Metal Mickey wants your job!

An interesting – worrying? – story this week about robots replacing humans. Foxconn, a Chinese company which makes products for Apple and Samsung (and possibly others) has replaced 60,000 workers with robots. Which probably isn’t great news, economically, if you work on a production line.

There’s a growing amount of speculation that the age of the robot may soon be upon us – here I cite in evidence the BBCs Will a robot take your job? article last year. But surely universities are safe from this kind of thing?

Not coming to a University near you

Don’t be so sure. Walking, talking, robots probably aren’t going to be lecturing any time soon, but a different sort of robot is having an impact. Georgia State University in the US of A has started using a chatbot – a computer programme which independently ‘talks’ to users via text – to handle queries from incoming students. Claiming to answer more than 99% of student queries, the chatbot, called Pounce, has been helping more than 3,000 students apply for financial aid; apply for accommodation and enrol in courses. In its first month of operation.

This is clearly a big thing. If the communication channels for conversion, admission and enrolment can be augmented by such technology, many universities will want to replicate this, and not only across the Atlantic. Speed and quality of response matter in converting an applicant to a student. And this directly impacts on universities’ success, so we can expect to see this grow. And if you have Pounce you don’t need so many people on the phones or at the help desks.

Pounce is a “product of artificial intelligence and supervised machine learning” according to the blurb from AdmitHub, its vendor. Its answering questions which cover a range of standard issues, but can come from left field:

everything from “When will I get my scholarship package?” to “Can my dog live in the dorm with me?”, according to the blurb. 

There are lots of other university situations where this sort of transaction happens. Exam time, submission of coursework; library and IT rules, graduation ceremonies. Universities are guiding students through all sorts of procedures, and a capacity to answer questions naturally, quickly, reliably and trackably, any time of day or night, is bound to be appealing to university managers. Where admission goes, other services will follow.

So it seems that, for some professional service staff at least, Metal Mickey might not want your job. But Hal from 2001 - A Space Odyssey might.

Context is everything

In an interesting debate on the Today programme this morning David Lammy argued strongly that contextual admissions, particularly by Russell Group universities, would be essential in helping to overcome structural inequalities in society. I thought some background might help to understand the argument.

Contextual admissions means the practice of taking into account, in deciding whether to offer a place at university to an applicant, factors other than the raw facts of a candidate’s exam performance. The logic being that an applicant who is getting ABB in A levels at a school with a very poor success record in exams has arguably demonstrated at least as much ability, studiousness and potential as a student who has got AAA at a school where 90% of students get AAA.

This is one of those arguments which appears differently depending on how close you are. If no-one in your school has done well at A levels previously, then you don’t have teachers who are practiced at coaching for A levels; you don’t have examples in the years above you of the habits of studying hard; and you don’t have peers within your class with whom you can study. I’d gladly subscribe to the abstract notion that if a student from this background has done really well, they’d have done even better had their background and school been different.

And now play the argument from the perspective of the AAA student who doesn’t get a place if the ABB student is admitted. The only direct comparison – the exam results – shows that one did better than the other. And the one who got ABB might well find it patronising to be treated on different grounds. Close up and personal, the issues can seem less clear cut: what is fair in the abstract has a whiff of rough justice when it’s two specific people.

That’s one reason why it’s been a controversial topic in education. Another is that entry grades are used by newspaper league tables as an indicator of quality. So the higher the average entry score, the better a university’s league table position.

I haven’t yet met an admissions tutor – in any kind of university – who doesn't want to take into account a student’s personal statement, or who doesn't want to recognise potential. A motivated student with slightly worse A levels is normally a much more interesting proposition than a high-achieving student who doesn't show any spark. The problem isn't that the staff in the ‘elite’ don’t recognise the issues.

So, if you accept that something should be done, what should that be?

Firstly, I think, take entry grades out of league tables. This would remove the perverse incentive to take ‘safe-but-potentially-uninteresting’ students. If you must have something there instead, calculate a value added score. Then you’re really measuring that universities do, rather than measuring, by proxy, the socio-economic characteristics of a university’s student population.

Secondly, get serious with improving all schools. There is a frightening disparity in achievement between state school pupils and those from private schools. Unless you believe that parental wealth correlates with the intelligence and ability of their offspring, this must relate in part to the quality and capacity of private schools. During the introduction of top-up fees in 2004, Charles Clarke, when accused of not being a socialist, retorted that if he was a real socialist he’d put all the money into pre-school education. And he has a point – if you’re going to solve social inequality by education, the earlier in someone’s life you start, the better. Affirmative action at age 16 plus is sticking plaster; a genuine cure should start years before.

Well, that's my two-penn'orth. I'll stand back now and watch the fireworks ...