On Wed, May 15, 2013 at 2:31 PM, Jed Rothwell <[email protected]> wrote:
> Cude apparently said: > >> Statistics have an important place, but I think this is sort of thing >> Rutherford was talking about when he said: if your experiment needs >> statistics, you should have done a better experiment. >> > This seems to be Cude's latest excuse to dismiss the research. Let me > reiterate: No, these experiments *do not* need statistics. You do not > need multiple experiments to prove the effect is real. One good one > suffices. > > One good one that can be performed by anyone, or for anyone. Such a thing doesn't exist, or it would have been done for the DOE panel. And when I asked for an experiment that one could do with an expected result, you linked to the Bayesian analysis. > Statistics are fun because, as Kevin O'Malley memorably put it: "It's not > that often that one can engage with someone who is demonstrably off by 4400 > orders of magnitude." > > Fun, except he did the math wrong. If you want to make simple arguments, at least get the math right.
