LETTERS:Interpreting statistics

From Mr J. Richard Pater Sir: Ten or twelve years ago I wrote to the Royal Statistical Society monthly newsletter, suggesting that our standard riposte to Lies, Damned Lies, and Statistics should be Fools, Damned Fools and non-Statisticians.

The problem is not, despite the oft-quoted line, that you can prove anything with statistics. It is that the opposite is the case: statistics don't prove anything. Descriptive statistics (unemployment figures, population census data and so on) are a picture - as reliable or unreliable as any other picture, but with beauty in the eye of the beholder. Analytical statistics, on the other hand, are about probabilities. They talk about degrees of uncertainty - rarely if ever about certainties.

Probability is a hard subject to grasp - look at the number of people who are prepared to gamble on the National Lottery. And even the pictures drawn by descriptive statistics can be hard to interpret. The job in both cases is best left to trained interpreters - like statisticians. It is not a job for fools, nor for damned fools, and certainly not for politicians.

Yours sincerely, J. RICHARD PATER Kendal, Cumbria 19 December