Main Interests Chatty Version

I study philosophy of science, specialising in philosophy of statistics. One way of explaining what this is is to contrast it with the fact that a lot of other philosophers study philosophy of probability, which asks questions like: what are probabilities? (Best said in a deep, resonant voice.) I’m magnanimous enough to think that’s a perfectly cromulent thing to study. But at the moment I’m more interested in how people use probability. So I study what statisticians do, and what other scientists do when they use probability, in as much detail as I can bear. This includes a large side-project on a rather ideological use of statistics called Evidence-Based Medicine which started in medicine but seems to have already spread to the rest of the health disciplines and to be spreading even beyond that.

The main problem I study is one which is surprisingly still an open question even though it has seemed very basic to everyone ever since the discipline of statistics took off a hundred years ago. Suppose you’ve got some data; you’ve got some hypotheses; then what? There’s still not even a rough outline of an uncontentious answer to the question of how the data can tell us anything about the hypotheses.

I haven’t got a complete solution but I’ve got a lot to say about categorising approaches. Why is that useful?

It’s useful because people — both philosophers and scientists — have been tending to focus on what I think are on the one hand one of the worst possible solutions and on the other hand one of the most subjectivist solutions, and have got rather stuck there. I can help by offering a way of (as it were) rotating the problem space to make it obvious that there are other possible solutions which are being ignored.

But at the moment, I’m mostly working on much less abstract stuff than that, questioning some of the obsessions of the Evidence-Based Medicine movement. This movement is under attack from several quarters, but it’s unusual for people to attack it on its own, technical, ground. This is what I call the “hit ’em where it hurts” strategy. Of course my work is not the vicious attack that that might suggest, and nor would such an attack be deserved; but a certain amount of criticism is deserved, because the movement is unnecessarily single-minded and simplistic, at least (I claim) from a statistical point of view and perhaps also (as many other people claim) from a more sociological point of view as well.

Jason Grossman written for the ANU School of Humanities Research Day, July 2007