Statistics: Don't let anyone do a number on you - Press Gazette

Statistics: Don't let anyone do a number on you

When identical triplets were born in January this year, the odds were quoted as 160,000 to one. Just three months later, the chances of having three identical babies had reportedly risen to 200 million to one.

Former Just Seventeen features editor, Wendy Varley, now mum of 21-year-old identical triplets, says the ‘ridiculous’figures so widely repeated in the media made her heart sink.

She says: ‘I groaned. I knew it had to be wrong. Identical triplets are rare, but not that rare. When they were younger I met other parents with identical triplets, some the same age as my own.

‘I know the population of the UK is somewhere over the 60 million mark, so if identical triplets really occurred once in 200,000,000 conceptions, you’d be talking about a once-in-many-generations event, when actually there are several sets born each year.

‘Using such a statistic is often a peg to hang a story on, but I don’t see the point if it’s wrong.”

Varley’s complaint is a familiar story. As a freelance parenting writer, I find it incredibly difficult to get to grips with working out the right statistics.

A mum may have faced difficulties conceiving or suffered a ‘rare’illness and gone on to have a family, and a news or features desk will want to know the odds of all of that happening. But the doctors who are prepared to talk to me, won’t be pressed. ‘I don’t do figures,’they say, resolutely – leaving me to have to work out for myself what can and can’t be said.

It’s a problem facing all journalists, whatever their specialism. The golden rule is treat all statistics with caution.

Freelance Kim Thomas says: ‘I would say avoid using statistics unless you can explain how they were arrived at and what they really mean.

‘The trouble is, sometimes an editor will ask you to do a story based on a statistic from a survey so you have to do it. You could take the stat apart and show why it’s nonsense but that would ruin the story.

‘The reason I believe most survey stats are nonsense is that they’re entirely self-serving – the survey has always been carried out by a company trying to sell something.”

Thomas adds that that most people – journalists and readers – don’t really understand statistics.

She says: ‘This particularly applies to statistics to do with risk. I mean, you can say to people ‘smoking increases your risk of getting lung cancer 10-fold’and they’ll still say ‘but my grandfather smoked and he lived to be 91″.  Tell them they’ve only got a one in 14 million chance of winning the lottery and they’ll say, ‘But someone’s got to win, haven’t they?”

Thomas has good advice for navigating a course through the maze of using stats.

She says: ‘I’ll always try and verify the statistic in some way, either by looking up more information or asking an interviewee how they arrived at the stat. If I think it’s really dodgy, I probably won’t include it.

‘I’d also say always look at the source. Look at whether proper procedure was carried out in the case of medical trials. Anything that comes from someone with something to sell or a PR firm is probably unreliable. If in doubt, ask an expert to explain it to you.”