On Mon, Nov 10, 2014 at 11:38 AM, Warren D Smith <warren.wds@gmail.com> wrote:
The chance that N statistical tests fail simultaneously =========Warren D. Smith====Nov 2014========
Suppose you perform N statistical tests. One fails with p-level a. Another fails with p-level b.
What is the p-level for this combined event?
Presumably the answer should not depend on what event we call "one event" and what event we call "another event". So the answer should be symmetric in the two arguments, that is, F_2(a, b) = F_(b,a)
It is NOT the case that F_2(a,b)=ab. Actually, F_2(a,b) = 2ab-aa = a(2b-a).
This is not symmetric. How do we tell whether the correct answer is 2ab - aa or 2ab - bb? Which event should we call "one event" and which should we call "another event" to obtain the correct answer? Andy -- Andy.Latto@pobox.com