Hi Rich,
We really need to identify reasonable prior probabilities in a
particular domain, agree on them as being the standard.
But how long do you want to wait for this to happen for each (new) domain?
Should people stop hypothesis testing until this happens?
For this reason I suggest to
Marcus,
I agree that we need agree on priors for a particular domain.
However, I don't agree with your alternative if I understand it
correctly. That is, it suffers from the same problem as Bonferroni in
that the posterior belief for a hypothesis is based on other
hypotheses evaluated with
Dear Marcus,
Thanks much for responding.
My real point was that Bonferroni (or any other correction) makes
little sense regardless of how we slice it. Consider these two scenarios:
1. Blood is found at the scene of a murder and the husband is the
suspect. They are about to test the
Hi Rich,
Since no other helping hand seems in sight, let me jump in:
I have 1,000,000 hypotheses that are not mutually exclusive.
1. I test them all. Do I apply the Bonferroni correction?
2. I plan to test them all, but I run out of resources after testing
only one of them. Do I apply the