In PR, good stats make good stories

You’ll have probably read last week’s other processed meat scare. The academic research made the front pages of most newspapers and websites. Everything from the BBC to the LA Times and, of course, the Daily Mail.

This is not a post on how to manage a scandal. Those have been done to death and fully paid-up member of Geeks Anonymous, like me, don’t do those.

If you didn’t see it, the short version is scientists have pooled questionnaire data from 23 centres in 10 European countries to reveal a strong correlation between eating lots of processed meat and dying earlier.

I know the story should have driven home the message of eating less junk, especially after the horse meat scandal, but I’m quite happy eating sausages etc – as long as the pigs are ethically reared, and brought up reading the Guardian, that is.

Instead, the key message for me was the dire need for good statistical analysis in PR. We have a lot of very bright people in this industry – just look at Steve Waddington or James Warren – but we seem to overlook anything bar the most basic statistical analysis.

As I said, I’m a geek. I learned to love stats during my PhD, I love Excel, I’ve even built client dashboards because, well, I was bored and had too much time on my hands (so shameful). So it’s not necessarily a big surprise that the desire for better stats in client news stories came to the front of my mind.

A quick straw poll of five friends working in PR – yes, I know this is completely unscientific – suggests we can easily calculate a mean, most of us know what a median is. But the terms chi-squared test (looks for variation in a population) and systematic review (pooling of all literature on a subject to create a larger, more accurate data set) escaped all bar one person.

We now live in a world where it has never been so easy to get hold of data. We can easily track down past surveys on any given subject and conduct a systematic review. We can use freedom of information to get hold of public data too – indeed many bodies put this online for us and UCAS is brilliant for this. And we can easily check it using a few basic statistical analysis tests.

One of the best ideas I’ve heard in a long time (sadly a friends, not mine) came in a recent beery chat about next year’s Commonwealth Games in Glasgow to predict how well Scotland might do.

By pooling data from past Olympics, World Championships and Commonwealth Games, it’s possible to get an idea of how the performance of the host nation improves and predict a medal haul, creating a headline like “Scotland to win X medals at Commonwealth Games according to analysis by sports drink manufacturer Lucozade.”

This associates the brand with the games, sports and success without (I feel) falling foul of the legal sponsor protection nonsense seen in last year’s Olympics. If only we’d have thought about it for last year’s Olympics.

That said, with both of us working for an electronics PR agency, it would have been harder to justify it for a microchip. Still, probably worth trying.

Rob Ashwell is an account director at Publitek Electronics PR.