Sunday 20 April 2014

Breaking the law of averages... statistically speaking

Well, this is a surprise – though I suppose it was always likely to happen one day. I’ve been invited by the Royal Statistical Society to enter for its Awards for Statistical Excellence in Journalism.
To be honest – a quality not always evident in the use of statistics – I didn’t know the awards existed. Or indeed the society, though I might have guessed such a thing was probable.
I’m flattered, of course, to be asked. My selection may, for all I know, be a matter of random chance, but as far as I’m aware none of my journalist friends has been invited.
So it seems a fair probability somebody at the august society has noticed that this column is a friend of statistics, used clearly and properly. And, even more, an enemy of statistics used improperly.
Too often they are used as a weapon to “blind people with science” – a tactic, much beloved of governments, which has a lot to do with blinding and stuff all to do with science.
Or just sloppily used by people who don’t know their statistical coccyx from their arithmetical humerus.
Statistics are incredibly useful, in many different ways. But they can also be incredibly misleading in the hands of people who don’t understand them – or who are banking on the likelihood that you don’t.
Opinion polls are one vastly over-used source of statistics, often mis-used, and frequently all but meaningless. Nevertheless, the following is rather illuminating.
What proportion of the national welfare budget do you think is spent on the unemployed?
According to a YouGov poll, people on average believe the figure is 41 per cent. In fact it’s 3pc.
And how much of it is fraudulently claimed by “scroungers”?
The popular belief averages out at 27pc. The reality is 0.7pc.
Why are so many people so wrong about these things? Because right-wing politicians, and their stooges in the media, have done such a good job over so many years of persuading us.
It’s a win-win tactic for them. By bringing the welfare state into disrepute they make it easier to dismantle it.
And by turning working people against each other they draw the flak away from themselves. It’s divide and rule in practice.
And here’s another statistic that says a lot about the degenerate state of British democracy.
The result of next year’s General Election won’t be determined by you and me, or anybody who votes in a “safe seat” – whether it’s safe for the Tories, Labour or Plaid Cymru.
It will be determined by the floating voters in a relative handful of marginal constituencies. Which, at a generous estimate, adds up to 4pc of the British electorate.
What those 4pc decide will have a major impact on so much – not least whether the NHS survives as anything like the life-saving, life-improving service we all know and cherish.
On the NHS, as on so much else, the present government has a cynical record of twisting the figures – though they have some way to go to match the US Republicans’ trashing of anything resembling proper health care.
But that is another story – one you can expect me to return to another time. When I do I shall aim to satisfy the Royal Statistical Society’s requirement to question, analyse and investigate the issues that affect society at large.
Over the course of my life I’ve been among the runners-up or on the short list for so many prizes and awards that the law of averages says I should win this one.
Except that the law of averages is a load of rubbish. If there really is such a thing, it surely states that most people who quote averages don’t understand them.
I heard some government flunkey the other day expressing outrage that “nearly half of all schools are below average”. Well, yes.
Make them better (assuming you actually have a proper way of rating them, which they don’t) and the average will rise. Leaving something like half (maybe the same half, maybe not) still below average.
And there’s another story there too.
 
****
 
And speaking of the misleading use of figures...
More than one paper reported recently that a warm day in Britain was “twice as hot as Athens!”
Someone should point out to them that 20 degrees Celsius is not twice as hot as 10 degrees – merely twice as far above the freezing-point of water.
It may make it clearer if it’s expressed in Fahrenheit. No one would imagine for a moment that 68 degrees was “twice as hot” as 50 degrees.
If any place on Earth was actually twice as hot, in absolute terms, as any other, then either one would be much too cold to support life or the other much too hot. Or both.
Between the coldness of absolute zero (minus 273 degrees C) and, say, the sun’s surface temperature (5,600 C), the range in which we can exist at all is really rather narrow.

No comments: