Multiple comparisons adjustments, null control etc. in bayestestR

[DominiqueMakowski]

In the debate whether yes or no Bayesian analyses could/should be "corrected" for multiple testing, comparisons and false positives, this suggests extending the null control method (implemented in JASP):

Null_Control_Jong.pdf

It would be nice to make something like that available in bayestestR

Could interest @mattansb

[mattansb]

These "adjustments" are for the prior odds between pairs/sets of hypotheses (not for any of the measures of evidence).
How would you like to see these implemented?

(Personally not a fan of these "let me tell you what your priors are" tools 😅)

[mattansb]

I don't think this is an avenue worth perusing (see previous comment about defining others' priors).

But if you are interested in such methods of default priors, you might also want to take a look at the BAS package, which computes term-wise adjusted priors.

[strengejacke]

(Personally not a fan of these "let me tell you what your priors are" tools 😅)

[bwiernik]

We do also have curve-wise intervals, so that is one method for multiple testing adjustment

[mattansb]

Yup, I'm def more in favor of those.
(The ones discussed here are for independent tests, or pairwise model-less tests...)