Lies, damned lies

(PHOTO: LUSI)

I signed up for a lunchtime seminar today called ‘Numbers and the News’ last week, and I confess that when the time came I almost skipped it. I wasn’t feeling exceptionally mentally agile today, and the idea of an hour of listening to a lecture about stats was not overly tempting. But I went, and I am so glad that I did.

It was given by the former presenters of a programme called More or Less and it was incredibly interesting. It’s become rather trendy these days for people to say ‘I don’t do maths’ and thus bow out of the responsibility of accuracy with numbers and figures. That’s not actually a terribly good idea, particularly when you are in any kind of journalistic or reporting career. As the seminar leaders said, you might as well say about poetry ‘I’m a scientist, I don’t do fiction’. So true.

They went on to explain how we all often fall victim to misunderstanding a phenomenon called regression to the mean. (There’s quite a good explanation here if you can ignore the irritating animated footprints). Basically, if you measure something at one point in time and get an extreme score, the next time you measure it it will probably have a less extreme score, for purely statistical reasons. So far so good, but why is that relavent? Because of cause and effect.

Muddling up the distinction between cause and effect and random co-incidence is an extremely easy mistake to make. And can potentially lead to wildly inaccurate reporting. Or to credit being taken where it isn’t necessarily due, such as in the effectiveness of speed cameras in reducing accidents. It’s so ubiqutious, that we were encouraged to look out for it whenever causes and effects are being bandied around.

It reminded me that while it can sometimes take me a bit of time to get your head around numbers, it’s not enough to bow out. The seminar was so entertaining that it was encouraging. Numbers rock.