Bayesian repeated measure analysis with post hoc Bayesian t-test, is it really wrong?
Hi EJ,
I have conducted BrANOVA with three groups and two conditions. I get a strong support for the interaction effect. I then follow up this interaction with (only) three highly hypothesis driven Bayesian t-tests. I really cannot see a problem with this, but you state in “Bayesian Inference for psychology. Part II” that there is multiple comparison problem. Is it really a problem when I am only interested in the interaction? The interaction is the only thing that is talked about and hypothesized in the study, nothing else.
Best,
Martin
Ps. The "Bayesian Inference for psychology" is great! Thanks!
Comments
Hi Martin,
A purely subjective Bayesian does not have a multiplicity problem. If the individual tests are "highly hypothesis driven", then I don't have a problem with it, especially for the given case with only three tests. But you can imagine that this will be an issue for, say, GWAS analyses or fMRI, where hundreds of thousands of tests are involved. Glad you liked the paper!
Cheers,
E.J.
Thought so, thanks for the fast answer!