frank_davis (frank_davis) wrote,
frank_davis
frank_davis

Correlation is Causation

In Sir Richard Doll's Times obituary, of his first paper - the London Hospitals study - it is reported that:

Doll thought that the increasing incidence of the disease might owe something to the hundreds of tonnes of tarmac being laid down across Britain at this time, but soon discovered that in 649 lung cancer cases there were only two non-smokers. Doll himself gave up the habit two thirds of the way through the research

Isn't that just the smoking gun? 647 out of 649 lung cancer patients were smokers! Pretty much an open-and-shut case of cause-and-effect.. No wonder Doll gave up smoking even before he'd completed the research.

At the weekend I was listening to an IPCC official insisting that there had been global warming over the past century. Of course, at the same time, carbon dioxide levels were steadily rising. Another open-and-shut case of cause-and-effect. No wonder the world is stampeding to shut down the carbon economy which has been creating our wealth killing us all for the past century.

A few days back Christopher Snowdon put up a chart showing that US oil production strongly correlated with rock music quality during the second half of the 20th century. Another great example of cause and effect. I guess that as more oil was produced, and America got wealthier, it produced more quality rock music celebrating that wealth. Something like that. It's obvious really. I'm surprised that I hadn't seen it myself.

And now today Letters From A Tory has - perhaps inadvertently - highlighted another one.

Even a basic Google search throws up evidence of impairments to visual functions, vigilance, reaction times, psychomotor skills, perception, drowsiness, attention and other cognitive tasks caused by alcohol, in addition to evidence of a clear relationship between blood alcohol levels and the probability of being involved in an accident. (my emphasis)

Clear relationship! That's shorthand for 'causal connection', of course. Clear relationship. Correlation. Causation. All the same thing.

That set me thinking about my own personal history of road accidents. The first one was when I came off my motorbike on the Thames Embankment in London about 35 years ago. I was annoyed at being sent on a futile, wasted journey across London. It rained while I was on my way back, and the roads became covered with a treacherous film of slime, and the bike skidded from under me when I angrily braked too hard. I didn't have a drink before. But I had one shortly afterwards.

Then there was the time, 10 years later, when I was driving a van, turning onto a road outside a pub car park, and the driver door swung open into the oncoming traffic, and as I reached out to close it, I forgot to take my foot off the gas pedal. The van continued turning, mounted the pavement and knocked down a lamp post. I must have drunk a whole half pint of beer just before that one.

And then there was the time I was driving my father back to hospital after a home visit. He'd had a stroke which had left him as strong as an ox, but mentally handicapped and unable to speak. Once he realised that he was going to be taken back to the hospital, he refused to get into the car. It took ages to induce him to get in. We'd hardly got a few hundred yards than he started to try to open the passenger door and get out. I went round a corner and clipped the front of a car emerging from a gate. I must've had a lunchtime glass of wine about an hour before that one.

The most bizarre one came after I'd been playing a Formula 1 Grand Prix computer game all day. This involved putting your foot down on the long curving straights, and then braking hard and turning sharply when you arrived at the next hairpin. After playing this all day, I jumped into my Mini to go buy something, and on the way back I found myself driving like I was in a F1 Grand Prix race. Arriving at a roundabout, I braked hard and turned the wheel over - and drove across the grass in the middle of roundabout. A few hundred yards later I spun the car through 90 degrees coming into another corner. And I hadn't touched a drop.

Of those four accidents, I put the first down to a combination of road conditions and being angry. And the second to a faulty door catch and my inexperience as a driver. And the third to the stress and distraction of my father in the car. And the last to having somehow had my driving brain temporarily rewired while playing a computer game. None were caused by alcohol. Alcohol was hardly a player at all.

And I've never had an accident driving while drunk. Not that I've ever driven while really totally plastered. But I'm sure I've been over the legal limit plenty of times. In my experience I drive perfectly well, technically, when I'm under the influence. I'm in control of the car. But I tend to get a bit headstrong. I'll take corners faster. I'll accelerate and brake harder. I'll throw the car around a bit. Driving seems too easy when I'm feeling a bit merry. And that seems to be the real danger: overconfidence.

When I drive I like to devote my whole attention to the road ahead. I hardly ever even play the radio. I don't like to even talk. And I like to be ice cool. Any distraction, and I drive worse. If anything is on my mind, and I'm worried or angry, I drive less well. Same if I'm tired or haven't slept well. Or if I'm driving an unfamiliar car whose handling I haven't figured out. It takes me months to get to know a new car.

In my personal experience, the real perils of driving arise when something takes my attention off the road ahead. It can be something that actually happens, like someone talking, or dropping something. Or it can just be something I'm thinking about, angrily or worriedly. If my attention is diverted for just a second or two, I can easily end up off the road.

But alcohol doesn't distract my attention. I think it slows up my reactions slightly. And it makes me a bit overconfident. But of itself that's not enough, in my view, to make me a menace as a drunk driver. There has to be something else to distract me if I'm going to end up wrapped around a tree.

But then, I've never driven a car with half of bottle of whisky inside me. So I don't actually know. Maybe I'll try it sometime.

I'm beginning to think tonight that maybe drink doesn't cause car accidents, but is simply a contributory factor. And maybe quite a small one. Far less than it's made out to be. And I already don't think that carbon dioxide causes global warming. It's just a contributory factor. And probably quite a small one. (I'm not even sure these days that the planet has been warming, given the poverty of the raw data, and the fact that it gets adjusted.) And I don't think smoking causes lung cancer. It looks more and more like a contributory factor. And maybe quite a small one.

It seems to be the besetting sin of our age to take small contributory factors and promote them to be prime causes. The strongest correlation points to the cause. Correlation is causation. The car crash happened because the driver was drunk. He got lung cancer because he smoked. The planet has been warming because CO2 levels have been rising. They are obvious candidates as causal agents, like prime suspects in a murder case. So much, in fact, that the police are so convinced that they will sometimes suppress counter-evidence in order to secure a conviction. Or climate scientists are so convinced that they will try to suppress sceptical points of view.

But I still reckon that US oil production was what made for all that great rock music of the 1960s. It's an open-and-shut case. It's obvious really.
Subscribe
  • Post a new comment

    Error

    default userpic

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.
  • 15 comments