Today the BBC have reported on the apparent positive effects on health of Transcendental Meditation:
"After nine years, the meditation group had a 47% reduction in deaths, heart attacks and strokes."Now, I suspect that this means that the there were 47% fewer deaths, heart attacks and strokes in the TM group than in the control group.
What the journalist has actually written is that quite a lot of them were dying, or having heart attacks or strokes before, and that's happening less now. Of course, if so many of them were dying at the start, there would be fewer people left to die or have heart attacks or strokes, so a reduction wouldn't be all that surprising.
On this occasion, this doesn't really make a great difference to anyone's interpretation of the story - it's obvious what the journo really means and we all learn that relaxation techniques are good for our health (I'd never have guessed).
But it's not always so harmless. During political conference season, I heard David Cameron on Radio 4 backing up his argument with the statistic along the lines of "there is a greater number of children living in the UK without a working parent than in other EU countries" (I can't remember the exact wording).
Sounds reasonable, and the journalist failed to challenge him. I say "failed", because he should have. There are only four countries in Europe that have a greater population than the UK - France, Germany, Russia and Turkey - and only two of those are in the EU. So it's not really surprising that there is a "greater number of children without a working parent" - most European countries have less than a fifth of the population of the UK.
Cameron's statistic, whilst possibly true, was utterly meaningless and should have been clarified or dismissed. Do journalists receive any training on this?
Thanks to Claire for spotting the BBC article.
[Edit: Just found this link - Statistics Help for Journalists - which assumes absolutely no prior knowledge and includes this gem of a quote: "Well, mathematicians have developed an entire field - statistics - dedicated to getting answers out of numbers."]