The truth behind statistics

More or Less Thursday, 27 July, 2006 BBC Radio 4, 1500 BST

According to a survey, 20% of people drink coffee while brushing their teeth.

Actually we made that up - though it'll be interesting to see if it catches on all the same.

In More or Less this week, you can hear more about all those surveys that "reveal" such fascinating - and often just as questionable - facts about us.

The World Cup, for example, added vigour to the nation's sex lives, according to one survey; and also ruined them as men preferred the football on TV, according to another.

We also continue to examine the way that statistical evidence is used in government policy making - or not, and conclude that quite often the evidence is as likely to give as much weight to the views of the bloke in the pub as to the standards expected by social science.

When trying to find out if something will work - be it a new policy, a new drug, a new teaching method - the gold standard of evidence is what comes out of a properly conducted random controlled trial.

How often to do we use this technique? In medicine, all the time.

Elsewhere, we use it only erratically. Without it, how easy is it to misinterpret the evidence? Using examples from street lighting and medicine, we investigate the commitment to finding out what really works.

Click here to have your say

Presenter: Andrew Dilnot
Producer: Michael Blastland

---

You can e-mail us by using the form below: