In article <[EMAIL PROTECTED]>, Rich Ulrich <[EMAIL PROTECTED]> writes > - I have comments on the question of corrections for multiple >testing. And I'm asking folks for feedback on Benjamini >and Hochberg's FDR as an alternative. > Not strictly relevant but there is also Holm-Bonferroni: Given N tests, go through the N tail probabilities smallest first, and stop at the rth test if its tail probability is > ALPHA/(N-r+1). Reject all null hypotheses seen before you stop. So the first probability gets a threshold equivalent to the Bonferroni correction, but the following ones get treated successively more leniently. Regardless of the number of false null hypothesis, the probability of rejecting any true null hypothesis is at most ALPHA, because that is the worst case probability of failing to stop at the first true null hypothesis. I had a URL for this which is now broken, so I'll leave you at the mercy of google. I think I saw somewhere that the original paper was getting pretty heavily cited.
Even less relevant is the following curiosity: Suppose that you do N tests and take the np - th tail probability. e.g. if p=1/2 you look at the median tail probability. If this probability is q, then the probability of observing a result of q or less is at most q/p. Because: if you pick a single probability at random from amongst the N, the probability that this result is <= q is q. If x is the probability that Np or more tests are < q, then we have q >= x * p + (1-x) * 0, so x <= q/p. The catch with this one is that you don't get to find out WHICH individual test is causal. -- A. G. McDowell . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
