Computing A-Level Statistics

Last night I tweeted this graph of the number of computer science degree entrants (2005–2011) and the number of students finishing A-Level computing over the same period (for non-UK readers: A-level is the exam taken directly before going to university):

I should state why this is interesting. There a couple of reasons why you might expect a link between the two:

  • There are various subjects, e.g. Maths, French, Physics, where an A-Level in the subject is required to enter the degree. In those subjects you would expect a causal relationship: while not everyone that takes French A-Level does a French degree, you would expect that a noticeable drop in French A-Level students would lead to a similar drop in degree entrants. For computing, this clearly is not the case, demonstrating that the causal link is not present (because computing degrees don’t require computing A-Level).
  • Subjects can sometimes have fads or overall changes in interest, e.g. the dot-com boom that affected computing in the early 2000s or the CSI effect on “forensic” courses in the mid 2000s. Since computing degrees do not require the A-Level, you would expect that any effect on the subject as a whole would be reflected in degree numbers too. But it doesn’t seem to be a whole subject decline. (Given the growth of smartphones and the web, it might actually be reasonable to expect a whole subject growth, but that doesn’t seem to be happening either.)

So the initial at-a-glance interpretation might be: computing is doing alright at universities, but in schools it’s in bad shape.

The UCAS issue

Obviously, both A-Level and degree numbers are affected by the available capacity, which is primarily determined by the number of institutions willing to offer those qualifications. However, universities cannot freely vary the number of students they have. The UK government does set limits on the amount of EU students that it will fund to take degrees per-university, and my understanding is that universities tend to dole out this allocation fairly statically to different departments. So rather than look at entrants, we could look at applicants — is it the case that applicants are declining too, and universities are maintaining their levels while taking lower-quality students:

Wow, that’s a fairly clear recent trend! That looks more like the subject growth I mentioned, and suggests that universities might actually be maintaining their levels while taking higher-quality students (generally, a larger pool to draw from means that universities can take the best students). It also makes the A-Level figures even more puzzling: A-Level continues to decline even while degree applications are climbing!

The ICT Effect

Some people on twitter responded to mention being put off the idea of computing by their ICT course. And indeed, longer-term, part of the reason for the low A-Level numbers is that some of the cohort clearly got diverted into ICT. To give an idea of the impact of ICT on computing: In 2003, Computing A-Level split into ICT and Computing. Before the split, 28,000 students took Computing. Immediately afterwards, 16,000 took ICT and 8,000 took Computing, dropping to 11,000 and 4,000 respectively by 2012. So the bulk clearly moved to ICT, but both subjects have since declined.

Maths Comparison

One interesting comparison is against maths. Maths is interesting because it can be taken as a bellwether for STEM subjects as a whole: it’s often thought that STEM is declining, so you might think the trend should be visible in maths (especially since so many STEM degrees rely on maths A-Level). In fact, Maths A-Level entrants went from 56,000 in 2003 to 85,000 in 2012. This provided me with a couple of headline statistics on twitter:

(All the figures are public data, taken from UCAS tables like this one for degrees — I used G4 for my statistics — and the JCQ results for A-Levels.

Let’s Fix This

A-Level Computing looks like it’s on the verge of dying out. This is not good news for the discipline as a whole — even though our degree numbers seem to be doing fine in spite of the A-Level decline, ultimately it would be good to see computing strong at all stages of the educational system. As it stands we face a sort of polarisation: those with computing degrees know computing, but almost no-one without a computing degree will have done any computing. (Compare to maths, where lots of students have maths A-Level, despite not doing a maths degree.)

Computing at GCSE has recently been revamped, and all the major exam boards have introduced a new computing GCSE for this September. It’s fairly clear that A-Level should get a similar revamp soon, to hopefully counteract any problems with the A-Level course content. The other change, which is slowly occurring, is that universities should start to value computing A-Level: as a minimum, computing degree courses should value computing A-Level, and an idealistic aim is that other subjects should value computing A-Level too. A big growth area in Biology is Bioinformatics: surely a degree in Bioinformatics should value (if not require!) computing A-Level. Similarly for the other sciences, where computer simulation is becoming a valuable or even vital tool in many sub-disciplines.

And we also need to continue our mission to make clear the difference between ICT and computing, so that students can make an informed decision as to which it is that they actually want to take. (Admittedly, by now many students have no choice: if their school/college doesn’t offer computing, it’s ICT or nothing — this also needs to be fixed.) If you’re a professional or teacher interested in these issues, think about joining or supporting the Computing At School group. We’re all working hard on exactly these problems.

4 thoughts on “Computing A-Level Statistics

  1. Well, we’ve had to put on another AS-level group after enrolment. The original 2 was based on applications for Computing with a ‘dropout’ factor applied. In numbers it’s about 25% extra (since group size is strictly limited by computers in room for practical classes this means we have to have another group). Still not enough girls though!

    This has raised staffing problems but ‘luckily’ IT has under-recruited which has helped.

    It will be interesting to see whether it’s a) the dropout factor being incorrect, b) students between applying and enrolling having got interested in Computing, c) increased awareness on the part of enrollers (‘Well, what about Computing as your fourth AS-level’) or d) some other reason altogether. I’ll do a bit of formal research in the MIS and bit of informal questioning when I see the students.

    It will be interesting to see how many of them stay the course, and also if the same is happening elsewhere.


  2. We have also had to put on another Computing AS group this year and lost an ICT AS group. We also have high numbers of our A2 class (20 students) applying for Computer Science related degrees.
    Howver, out of a cohort of 50 this year, not one is a girl. There is one girl in the A2 class. I wonder how much this is relected in degree applications for the course?

    1. A rough rule of thumb for university CS is that around 10-15% of students are female. Glancing at the statistics, it was under 10% female for A-Levels in 2011, but degree accepts were around 15% female. So looks like university is doing slightly better than A-Level.

Leave a Reply

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

You are commenting using your 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