Thursday, 20 October 2011

Damn Statistics

I like graphs and charts. It's part of my job to use them to communicate data effectively. I try to do so clearly and fairly.. because there's usually nothing to be gained from hiding from the truth.

Eoin Clarke has done a chart here to show why Independent schools are much worse than comprehensive schools. As we can see, their GSCE results are much worse.

Oh.. but wait.. they're not. They're very slightly worse.. but he's used a deeply selective and misleading scale to make a difference of 0.4% look huge.

I did a chart using the same source data. But I added a second set, which shows the comparable GCSE results for the schools, but for pupils who got 5+ GCSE passes including maths and English. I'm not saying that this is a more telling indicator of quality.. that's for the reader to decide.



I've also used a selective and misleading scale.

Aren't comprehensive schools RUBBISH!

Here's another chart.. with a proper scale.. that shows the relative performance differences according to whether one thinks that Eoin has chosen the best measure, or whether ones thinks that we should take into account whether or not pupils can read, write and do sums.

1 comment:

  1. The Thought Gang,

    I need a favour ( a small one).

    If I type into my excel sheet the following

    a) 7.3
    b) 8.1

    And create a bar chart.

    The bar chart is ugly and out of scale.

    Its one of the reasons I make the numbers above so big (see the graph u refer to for example).

    I genuinely wish for the graph to show the correct scale but excel has got it into its head that their close up scale is preferable.

    My email address is eclarke04@qub.ac.uk

    Could you email me instructions as to how best I can rectify this?

    Kind regards,
    Éoin

    ReplyDelete