At 12:29 14/09/98 +0100, Ted wrote:
snip>
>A: Is there any real evidence that false-positive rates exceed those
> implied by the P-values?
>
Ted i dont know of direct evidence, but multiple testing can very quickly
inflate the "true" p value way over the nominal p value of 0.05.
if the trial/experiment is repeated by other investigators, this should
help control the problem: from personal experience in lab stuff i strongly
suspect this to be the cause of some "unrepeatable" pieces of work. for
this correction to be applied of course implies that you can publish it.....
regards
========================================================
"You can't step in the same river twice." -- Heraclitus
"You can't even do it once." -- Heisenberg
========================================================
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|