In From Eros to Gaia, Freeman Dyson criticizes big scientific research projects with huge price tags.His argument is simple: by the time these big things get built, the science that justified them has shifted. They aren't quite what is needed, but, given the enormous amount of money already spent, they have to be built anyway. He was thinking about major astronomy projects, mainly, but it seems that his warning applies also to major public health studies. Today's papers reported that a very large (and so not repeatable) study has shown that there is no connection between a low-fat diet and decreased risk of heart attacks and cancer. Critics charge that the study did not consider a low enough fat target, that it did not distinguish among kinds of fat (as recent research suggests it should have done) and that it did not consider the effect of combining a low fat diet with regular exercise.
One can't help thinking that this big, conclusive, never to be repeated study was done way too early, and that it was way too big and too long. It answered conclusively a question that many researchers have come to think is just the wrong question to ask, and it suggested to lots of people a result quite different from the one it actually established, "Don't worry about dietary fat."
One sees a place for good screening of big, expensive science -- and for a kind of research conservatism.
Beyond that, I think this story has some lessons about major research efforts and conclusive refutations in general. They are always tempting: one wants to do one's best, to get it really right -- for all sorts of "its." But there will be revenge effects for almost any major research effort, and for almost any conclusive proof. One's over-all epistemological situation may be better when one rests satisfied with "pretty good reason," especially if one has "pretty good reason" to believe lots of things and if one revisits one's "pretty good reasons" pretty frequently.
Posted by shea0017 at February 8, 2006 9:30 AM