# View Article

view the latest news articles
Friday, October 26th 2012, 1:23 PM EDT
I had to pinch myself over this story from Graeme Archer at The Telegraph.....The L'Aquila earthquake trial reminds us that scientific evidence shouldn't determine public policy!..Graeme have you not thought of another area this may apply?

The L'Aquila earthquake trial reminds us that scientific evidence shouldn't determine public policy! by Graeme Archer - The Telegraph

People find it hard to understand the nature of risk: discuss. A pertinent assertion, in the week that scientists have been found guilty in an Italian court for understating the likelihood of the L’Aquila earthquake. Moreover, the assertion is true, as can easily be demonstrated. Stand behind someone in the queue at WH Smiths while they purchase a lottery ticket, and watch the care with which they select “their” (irrelevant) numbers. Or travel across the Atlantic on a plane, sat beside me.

The former is less physically demanding, as I’m less likely to claw at your arm in terror during the purchase of a lottery ticket than I am while the plane bounces around in turbulence. There’s no point telling me that there’s a very low probability of falling from the sky in a ball of flame, that such disasters happen only rarely. I don’t care about “long-run” arguments: I care about this flight. If the probability of any event is non-zero – if there’s a finite chance that it will occur – then it will happen, at some point; and nothing in the construction of a long-run probability (in its English sense, or its precise mathematical expression) has anything useful to say about any particular instance of any particular flight.

Neither the lottery ticket purchaser, nor the statistician who hates flying, is entirely irrational, however. What matters when making a decision isn't (just) the chance of a good or bad outcome; it's the expected cost (or gain) of those outcomes, after deducting the price of the stake.

It costs a quid to buy a lottery ticket, so while you've a vanishingly small chance of winning, your cost of losing (albeit with high probability) is small. And if you didn't accept my argument that even unlikely events must eventually happen at least once, then you don't understand the popularity of the lottery.

Similarly, while the "cost" to me of flying is a lot more, in psychological terms, than a quid – it takes six hours to fly from London to Philadelphia, which in fear terms means “forever” – I do still fly for work when I must, because the price of not flying would be to have to find another job. So the expected cost to me – whatever the fear costs, multiplied by the chance of falling from the sky in a ball of etc etc – must be less than the cost of having to find another job (multiplied by what my head, if not my stomach, knows to be the chance of making it to Philly alive).

Which explains why estimating the chance of an earthquake is such a high-stakes game; or rather, it explains the difficulty in deciding what to do, in light of that estimate of risk. It’s no use saying “the chance of a quake is low (or high)”, and stop there, because the cost of taking no action if the quake occurs vastly outweighs the cost of taking action, only to find the prediction was wrong. It’s the dilemma expressed most venally in Jaws; remember how the mayor discounts the cost of a shark eating his electorate because he thinks the chances of it occurring are not as high as Rod Steiger keeps asserting?

The Italian scientists argued correctly that their science is by its nature imprecise (at least we can state with precision the probability to win the lottery; such precision is impossible in geophysics) and that their deliberations could only be the precursor to political decisions, taken by those charged with weighing the costs of acting or not acting. The chance of an earthquake being low does not mean that one will not occur tomorrow.

I suspect the judge will find his decision revisited by the appeals process, and so doubt if the wave of panic from Establishment Science is justified. As Prof Malcolm Sperrin of the Royal Berkshire Hospital, said: “If the scientific community is to be penalised for making predictions that turn out to be incorrect, or for not accurately predicting an event that subsequently occurs, then scientific endeavour will be restricted to certainties only and the benefits that are associated with findings from medicine to physics will be stalled.”

Of course he is right. But a possible good outcome from this unwelcome court ruling might still follow. In recent years we’ve seen a combination of two unwelcome phenomena: politicians trying to avoid responsibility for their decisions, by hiving them off to independent bodies of “experts”, and simultaneously some “experts” trying to supplant the role of politicians, by asserting that their “evidence” must outweigh any political judgment. Remember Prof David Nutt and the Home Secretary in the last government, arguing about whether or not drugs policy should be altered?

Experts must remember that the estimation of risk is the first step in a process, which then proceeds to use those estimates to evaluate the cost of various political actions. Political decisions have consequences which are impossible to determine accurately; something scientists – now understandably clear that they don’t wish to pay the price for making the wrong call – should remember. L’Aquila was a tragedy for which the physicists weren’t to blame. But who’s to blame for thinking that scientific evidence should determine policy? Some decisions have expected costs that far outweigh anything so trivial as the lottery.