Statisticians, he said, view journalists as innumerate. They are unable to understand quantitative reasoning, are prone to distort and oversimplify, and they won't listen to anything you have to say. Journalists, on the other hand, perceive statisticians as illiterate pedants, who couch their answers in so many ifs and buts that they lose the message of the story they're trying to convey. And they won't listen either.
These are broad-brush stereotypes, of course, and they certainly don't apply to all journalists and statisticians. Indeed, McConway was at the Royal Statistical Society International Conference today to explain his work with the Science Media Centre (SMC), a group set up to be a rapid-response vehicle to connect academics and journalists in order to comment on and critique science news before it hits the headlines.
The SMC also invited along two journalists – former ITV News science and medical editor Lawrence McGinty, and BBC science writer Jonathan Webb – to put across the media perspective, in a bid to reach a mutual understanding of what each group needs from the other.
McGinty and Webb both welcomed (indeed, invited) help from the statistical community to help them do their job. 'Please work with us,' Webb said. 'Don't just carp about what we do when we get something wrong.'
But the message they wanted to stress was the importance of timely responses to requests for help. 'News works in a very different universe to research', Webb explained. 'If we ask for a response by 4pm, that doesn't mean "4pm if you have time, but otherwise tomorrow will be ok". It means 4pm.'
Often the time pressure is worse than that, said McGinty. The advent of social media means 'the deadline is now'. Public relations people know this, McGinty said, and they try to 'take advantage of that' – to frame a story in a certain way.
This is where the SMC is designed to play its role, swooping in to stop the spin before it happens, or to challenge it after the fact.
Patrick Wolfe, professor of statistics at University College London, also works with the SMC. He highlighted key points 'worth remembering' for other statisticians looking to join him.
'Statistical literacy is key', he said. Statisticians need to be comfortable explaining a p-value to a journalist in under 30 seconds, for example.
He also asked statisticians to remember that, when you work with a journalist, 'you are there to explain, not peer review'. A journalist might be basing an article on a journal article you find questionable, but 'the best you can do is explain in an accessible way, what are the positives and negatives, the caveats, and how seriously to take the conclusions', Wolfe said.
