Go Figure: Watching out for Wimbledon-washing machine links
What's the link between tennis on TV and washing machines? If you suspect a weird connection, ask a statistician, says Michael Blastland in his regular column.
"Do you repair washing machines?"
"Indeed we do. Wimbledon is it?
"The tennis. Been watching?"
"Er, a little. Don't tell me… you're a tennis-lovers-only washing-machine repair shop."
"No. Just that we were quiet for Wimbledon, booked-up now."
"Really? Wimbledon affects demand for washing-machine repairs?"
"What is it, two weeks' desire for whiter-than-white whites that wears them out?"
"Could be. Or that people don't wash anything because they're glued to the telly."
Let's stop there. I love weird links and explanations. They remind us of the difficulty of knowing for sure what causes what.
Of course, it could be just the usual story of people spending less when they're preoccupied, much like "Olympics/World Cup-watching costs UK economy billions", as headlines like to say.
But the fact the washing machine shop seems to be having a post-Wimbledon bounce shows what's often wrong with "the cost of…" stories. They fail to take account of catching up later. Cause and effect can be more tangled than portrayed.
So maybe washing machines have nothing to do with tennis. It is a bit surprising that people apparently go without one for up to two weeks, watching with enough devotion to make the difference between no business at all and booked solid.
Well, it is to me anyway. So maybe it's not Wimbledon what done it, maybe it's Glastonbury mud. Maybe both. Maybe it's the weather.
Someone somewhere in Britain's statistical machinery will be counting the business done by washing machine repair shops. They will, really. Business surveys even count turnover in kebab vans.
And when there's talk of a return to recession, the explanation for a fall in business will matter. One such explanation earlier this year was snow. Some economists now begin to wonder if that was the whole truth.
Part of the problem here is our ability to tell stories to fit any fact. Let's say we hypothesise that it's not Wimbledon, it's the weather, all that rain. And then we remember that this year's Wimbledon was mostly sunny.
Oops. But never mind, because now can we say that hot weather makes people sweaty, just as rainy weather makes them soggy. Either way, more washing and more breakdowns.
There's danger as well as genius in the ease with which people construct graceful narrative arcs from any two things that grab their attention, like tennis on telly and the business bottom line.
So who knows the real explanation? Who knows if it's generally true that there's a repair slump during Wimbledon fortnight or if this is one shop where whoever should be on the phone dreams instead of Andy Murray.
Ok, it is only a washing machine. Apart from me running low on socks, should you care?
Kind of - if you care about how we claim to know stuff. The point of all this is to argue that it helps to have a restless imagination.
The book Freakonomics became a hit partly with a controversial argument that falling crime was due not to law enforcement but because abortion became easier in the 1970s, meaning fewer potential criminals were born. Whatever you think of that, it's good to ask ourselves what we might previously have missed.
Which is why I prefer stats to politics. Doing good stats means exercising a pathological interest in the story that might have been missed. Doing politics can seem to mean a near pathological interest in telling us why your story was right all along.
So back to tennis and washing machines, what do you reckon? Was there a repair slump? Is Wimbledon the explanation? If not, what is? And how would you find out?
Answers welcome in the usual way. The best, by which I mean the most ludicrous as well as the most plausible, will go in the readers' comments below.
Mind you, do I have any robust evidence for this difference between politics and stats? None whatsoever.
It's all anecdote and prejudice and not even the opinion of the bloke in the washing machine shop. So how would we find the evidence? Random controlled lab tests on ministers and professors of statistics, maybe? More answers please.