Why might black students be under-represented at Oxbridge?

Equality and diversity at tertiary education level are  a tangled set of important and emotive issues. Needless to say, David Lammy has made a complete dog whistling breakfast of the whole thing. (Note: all the following figures not taken from his article are from table 5 in this summary, and relate to home students who filled in an ethnic diversity form.)

In an expose of racism in Oxbridge admissions for 2009 (on the watch of the last Labour Minister for Higher Education, apparently a Mr David Lammy) he has perpetrated statistical no-nos such as:

  • generalising from insignificant sample sizes (highlighting 1 Black Caribbean admission in Oxford for 2009, but failing to mention that it comes from 35 applications in that category, which in 10,210 applications and 2,653 acceptances, is pretty much noise),
  • cherry-picking (choosing the Black Caribbean group at all, with its acceptance “rate” of 2.9%, as opposed to, say, the Black Other group with an acceptance rate of 21.4% – but more about that anon),
  • a particularly technical error which I think is known as “making stuff up” (stating that there are no Black faculty members at Cambridge, something which will come as an enormous existential surprise to people who are in fact Black faculty members at Cambridge).

This is very sad, because a look at Table 5 suggests there is potentially something to investigate here. That healthy-ish 21.4% success rate for Black Other students? Also an inappropriate cherry-pick, of course. Hey, I could be a shadow minister! It represents 3 admissions out of 14 applications. Following Lammy’s logic, we’d have to surmise that some highly complex form of racism was going on which selected Black Other applicants over Black Caribbean applicants. This is clearly unlikely to be the case.

Anyway, averaging out these two groups along with Black African (which I realise is a bit of a lumpen way to treat people’s ethnicity, but they’re pretty lumpen categories anyway) you get 27 acceptances out of 221 applications, or a success rate of 12.2%. That sample size, 2.1% of applications and 1% of acceptances is still within the margin of error, but let’s be generous and take it as an indication, if nothing else. Not at all healthy, is it, given that the overall success rate for all British domiciled students is 26%.

If this indicator does represent a genuine problem, and not statistical noise, what might be happening? This is Oxford’s gloss:

Black students apply disproportionately for the most oversubscribed subjects, contributing to a lower than average success rate for the group as a whole: 44% of all black applicants apply for Oxford’s three most oversubscribed subjects, compared with just 17% of all white applicants. That means nearly half of black applicants are applying for the same three subjects … the three toughest subjects to get places in. Those subjects are economics and management, medicine, and maths… This goes a very long way towards explaining the group’s overall lower success rate.

(Note: Where I have placed the second ellipsis, the Guardian has the words “with 7% of white applicants.”, no capitalisation, straight after the full stop of the previous sentence. I have assumed this is some kind of subbing error, but can’t be sure. They spelt “rein” as in “to rein in” with a g on their front page the other day, so anything’s possible.)

Now, this could well be special pleading on Oxford’s part. We just wouldn’t know without seeing the full figures, which they’re not releasing any more than Lammy is. Frankly, the way Lammy appears to have mauled the stats, I don’t think Oxford could do worse than release the lot in full – ethnicity breakdowns by subject and college.

But, pending such mere inputs as hard data, the Oxford response has a slight ring of truth about it, and I’ll tell you why. We might posit that as a general rule, black kids are less likely to be applying from public schools, private schools and top state schools. They are disproportionately educated in lower-achieving schools in poorer areas. That means that, amongst many other unfortunate things, their Oxbridge preparation is not going to be so hot. And one of the things they tell you in Oxbridge preparation is “maximise your chances”. If you want to apply for a big name subject, but there’s another less popular variation that will do just as well, go for that. If you’re not that fussed about golden twiddly bits on your college, go for one of the concretes or redbricks as your first choice.

This is how Oxford (and broadly Cambridge too, as far as I know) admissions works: you pick a first choice college, and you get allotted another two (alternatively, you can make an open application, in which case you’ll be allotted three colleges, and they will probably be the ones with fewest applications). If you’re called for interview, your first choice college will interview you. If they want you, that’s it, no more interviews. If they don’t want you, and your second and third allotted colleges still have places to fill, you’ll be sent along to those. And here’s the crucial thing: all of this happens within a week.

You can see what happens to people who apply to the popular colleges. Not only have they lowered their chances by applying to a competitive place as their first choice – their second and third choices are compromised too. Because by the time their first, popular choice has chewed them up and spat them out, the second and third colleges have already filled plenty of places from their own first choice (or allotted open) applicants. Fewer places remain for the scramble of applicants rejected from popular colleges. And if you’ve picked a popular subject as well, then you’ve exacerbated the problem. Very probably, this means that some of those who get rejected every year could have got in if they’d applied to a less popular college, or subject.

The point of all this is not, of course, to imply that black students can’t get onto these courses, or into popular colleges. All other things being equal, they stand the same chance as everyone else – but everyone’s chance is lower than the chances of those applying for less popular subjects and colleges. So a disproportionate number of black applicants to popular subjects would indeed have the effect the spokesperson suggests.

But, as I say, this isn’t really demonstrable without the full figures *hint to media*. As a matter of fact, the most shocking statistic for me to emerge from the whole thing doesn’t have anything to do with Oxbridge admissions. It’s this one:

In 2009, more than 29,000 white students got three As or better at A-level (excluding general studies) and about 28.4% applied to Oxford; while 452 black students got three As or better, and nearly half applied to Oxford.

*attempts Steve McQueen impression* Four hundred and fifty two? There are probably more people than that in John Lewis on Oxford Street right now. They are outnumbered by our MPs. Jesus.


  1. Since black people are about 2% of the population, 452 with 3 As or better compared with 29,000 whites looks about right to me, even good, if you take into account the concentration of black people in the lower income groups. So why the surprise?

    That special knowledge is required to get into Oxbridge, is one of the ways in which upper income groups and private schools are privileged, and one of the aspects of Oxbridge entrance that needs reforming. We also need to move away from a society where a third in ancient Mesopotamian from Oxbridge opens more doors than a first class degree in civil engineering from anywhere else.

    Even if there were some factual errors in Lammy’s article he highlighted a problem which undoubtedly exists. Instead of making excuses Oxford and Cambridge need to be seen to be tackling it.

  2. Ha, I stand corrected on the A-level results. Should have done that sum.

    On the “special knowledge”, I think that’s rather my point. It’s a disadvantage that desperately needs to be addressed by better information – and the responsibility for disseminating that does lie broadly with Oxbridge outreach programmes. But I think that’s what I object to in Lammy’s account. He’s not looking for the *mechanisms* by which black students are under-represented, and thinking about ways to counteract them – he’s blaming the problem on bigotry/complacency/the old boy network/insert cliche here and his solution is pretty much along the lines that Oxbridge need to buck their ideas up. That doesn’t really do a lot for the kind of problem I’ve raised (assuming I’m right about it). So I don’t think that Oxford’s spokesperson offering some pretty pertinent numbers (subject to further analysis) should be read as them “making excuses”. If what they say is true and verifiable, they’ve offered a potentially more fruitful approach to this than Lammy has.

    None of this precludes that individuals involved in selection at Oxbridge might be culpable of course. There’s a broad spectrum of attitudes that could contribute to this problem, and it’s not necessary for someone to be a raving bigot to nonetheless be under-aware about, say, cultural difference and how to read it. I could quite believe that this exists.

    Having said all of that, I think it has yet to be demonstrated statistically that there is a problem, at least in the case of the Oxford 2009 stats, which is all I’ve tried looking for in the public domain to match against Lammy’s figures. 2% of the population, 2% of the applications, 1% of the admissions is within margins of error. It certainly doesn’t rule out there being an problem. Neither does it prove one. It invites further analysis, going back several years. In which case, again, it would be great if Lammy would release the full figures he’s garnered.

  3. Just to fill in the detail on how Cambridge applications work: It’s much the same, but without the two lesser preferences – you just apply to a particular college, or make a pool application (and as you say, people who get pooled are quite likely to be fished out of the pool by the more under-subscribed colleges). An application to an individual college will have one of three outcomes: acceptance, rejection, or being pooled anyway. This has the advantage that you don’t need to arrange to be around for the week, in case someone else wants to interview you.

  4. I was speaking to someone involved in applications at Cambridge, and he told me that there is a bit of reverse social engineering going on, entirely unsuspectingly. Apparently, if someone from a background traditionally viewed as privileged (eg a white public schoolboy) crops up in the pool, they are often picked up early *not* because tutors are all prejudiced, but because there is so much noise about the low intake of minority groups and people from poorer backgrounds that they assume white public schoolboys won’t get in at another college. As opposed to someone from a working class background or a black african, who they assume will be snapped up. Now, that’s only anecdotal, obviously, but it makes a certain amount of sense. I am more ready to believe that oxbrige tutors are broadly sympathetic people than that they’re all racist snobs (which in my experience couldn’t be further from the truth).

    But what would fairness look like, anyway? Surely not simply modelling on the ethnicity proportions of the UK, which a) would reveal it as nothing more than a cosmetic exercise and b) would only work if Oxbridge didn’t take overseas students, which they do.

  5. You choose to highlight the fact that 452 black students get 3 As at A level compared to 29,000 white students. OK, that clearly indicates that black students are underperforming whites in this category – whites are prima facie perhaps 1.5 times more likely to be in this category than blacks (though that’s based on the overall percentage of the UK population who are black – the percentage of the relevant age group is certainly higher). Now compare the Oxford stats that you yourself give: 35 out of 10,021 applicants to Oxford are Black Caribbean. That suggests that Black Carribbean people (about half of the black population of the UK) are applying at something like a third of the rate of non-Black Carribbeans, whereas their acceptance rate of 1/35 compares pretty unfavourably with the 1/5 acceptance rate in the general applicant pool, even accounting for the confounding factors mentioned. There are good reasons moreover to focus on Black Carribbean as a group rather than Black Other: Black Carribbean correlates in the UK to a specific disadvantaged social group, whereas Black Other can cover members of the ruling classes of African nations educated at British public schools for example.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s