Physics and Statistics

Predictably, yesterday’s newspapers and other media  were full of feeble articles about the A-level results, and I don’t just mean the gratuitous pictures of pretty girls opening envelopes and/or jumping in the air.  I’ve never met a journalist who understood the concept of statistical significance, which seems to account for the way they feel able to write whatever they like about any numbers that happen to be newsworthy without feeling constrained by mathematical common-sense.  Sometimes it’s the ridiculous over-interpretation of opinion polls (which usually have a sampling uncertainty of ±3 %), sometimes its league tables. This time it’s the number of students getting the top grades at A-level.

The BBC, for example, made a lot of fuss about the fall in the % of A and A* A-level grades, to  26.3% this year from 26.6% last year. Anyone with a modicum of statistical knowledge would know, however, that whether this drop means anything at all depends on how many results were involved: the sampling uncertainty depends on size N approximately as √N. For a cohort of 300000 this turns into a percentage uncertainty of about 0.57%, which is about twice as large as the reported fall.  The result is therefore “in the noise” – in the sense that there’s no evidence that it was actually harder to get a high grade this year compared with last year – but that didn’t prove a barrier to those editors intent on filling their newspapers and websites with meaningless guff.

Almost hidden among the bilge was an interesting snippet about Physics. It seems that the number of students taking Physics A-level this year has exceeded 35,000 in 2013.  That was set as a government target for 2014, so it has been reached a year early.  The difference between the number that took Physics this year (35,569) and those who took it in 2006 (27,368) is certainly significant. Whether this is the so-called Brian Cox effect or something else, it’s very good news for the future health of the subject.

On the other hand, the proportion of female Physics students remains around 20%. Over the last three years the proportion has been 20.8%, 21.3% and 20.6% so numerically this year is down on last year, but the real message in these figures is that despite strenuous efforts to increase this fraction, there is no significant change.

As I write I’m formally still on Clearing business, sitting beside the telephone in case anyone needs to talk to me. However, at close of play yesterday the School of Mathematical and Physical Sciences had exceeded its recruitment target by quite a healthy margin.  We’re still open for Clearing, though, as our recent expansion means we can take a few more suitably qualified students. Physics and Astronomy did particularly well, and we’re set to welcome our biggest-ever intake into the first year in September 2013. I’m really looking forward to meeting them all.

While I’m on about statistics, here’s another thing. When I was poring over this year’s NSS results, I noticed that only 39 Physics departments appeared in the survey. When I last counted them there were 115 universities in the UK. This number doesn’t include about 50 colleges and other forms of higher education institutions which are also sometimes included in lists of universities. Anyway, my point is that at most about a third of British universities have a physics department.

Now that is a shocking statistic…

Advertisements

7 Responses to “Physics and Statistics”

  1. Bryn Jones Says:

    Oh, yes. I regularly wince over the over-interpretation of statistics by the British media, particularly the drawing of conclusions by neglecting errors (statistical uncertainties).

  2. To offer more damn lies…

    The physics figures for the nations (Wales and NI in this context) are interesting in that they both show dramatic falls in the number of girls choosing A-level physics and even in England the numbers rose only a little, albeit in a reduced cohort. In fact, the A-level results as a whole show a tendency towards reinforcement of gender stereotyping in many subjects.

    By the way, depending slightly on how you count them, there are now around 50 universities in the UK offering physics degrees that are either accreditied or likely to be and several others (e.g. UEA and Reading) that offer degree coursess with a great deal of physics in them.

    • telescoper Says:

      I also now realise that some departments might not be in NSS because they didn’t get enough students to respond; there’s a minimum rate of 50%.

  3. I’d also like to see uncertainty reported in measures of economic growth and unemployment.

  4. I remember during the flu ‘epidemic’ a couple of years ago, I scribbled down some basic estimates of noise in the surveys and found much the same: the BBC was reporting noise. It’s depressing, but I doubt there’s much to be done.

  5. […] Physics and Statistics (telescoper.wordpress.com) […]

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

%d bloggers like this: