When science has no chance

An enlightening piece on the 2012 conviction of Italian seismologists by David Wolman.

The prosecutor offered, and the judge accepted, the theory that Italy’s leading seismologists had conspired with the mayor’s office in l’Aquila to send a falsely reassuring message to the town’s residents, who had been warned of an impending disaster by a lay forecaster called Gianpaolo Giuliani, a charlatan by the scientists’ standards.

The scientists said the likelihood of an earthquake was “low”. That meant two things: first, even in a high-risk area like l’Aquila, a major quake was unlikely to occur on any given day; second, the likelihood of it happening within a short-term period was no higher on the day before it did than on any other day.

Unfortunately, some residents latched on to that “low” probability and believed the earthquake was “unlikely”. That’s why they did not leave their homes when the first shocks hit, even though normally they would do just that.

The scientists’ statement, in essence, was: “You should not worry today about tomorrow any more than you worried yesterday or the day before yesterday. This is not to say you should not worry at all – do not forget that you live in a high-risk area. There’s always a risk, although relatively small, but it’s the same every day. When Giuliani tells you that you should worry more today than you did yesterday because of some gases or energy escaping, that’s rubbish.” But it was never communicated properly. The people of l’Aquila were confused and disaster ensued.

At the trial, the prosecutor claimed the scientists had lied: they knew the probability was high, not low. He cited a map produced by the government research center where two of the defendants had worked showing probabilities of a major earthquake occurring within 50 years. The province of Abruzzo and l’Aquila were in a high-risk zone. Bingo! No, not really. In absolute terms the probability of an earthquake happening in l’Aquila within several days was still low, and equally low on any given day. It was only high relative to other places — but the scientists were not warning people who planned to move to l’Aquila from elsewhere. Their audience were l’Aquila residents, for whom this risk was part of everyday life.

The prosecutor also cited a paper published by one of the defendants the 1990s with a hypothetical predictive model which resulted in the probability of an earthquake in Aquila in the next year being 100%. If the model had been correct, there should have been an earthquake there within a year. But there was none. As another scientist explained on the stand, the model proved itself wrong – in a self-evident [and possibly embarrassing] way – and no reasonable scientist, including the author of the paper, would consider that model valid thereafter.

The judge, however, claimed in his written opinion that the author of the paper continued to believe in his own discredited model. That sort of idiocy, as I have learned, is typical of Italian courts. The point of the trial was, apparently, not to find justice but to provide some sort of consolation-through-retribution to the victims of the latest l’Aquila quake – and to humiliate the bespectacled.

3 Comments

Leave a Reply

Your email address will not be published. Required fields are marked *