Bayes Factor comparison for multiple datasets

[Jan Hackel]

Hi RevBayes community,

I have a question perhaps more related to Bayesian model comparison in general than RevBayes itself:

I want to compare the fit of biogeographic models that differ in manual constraints/dispersal multipliers, but not in free parameters.

1. Am I right to assume I can use Bayes Factors for this kind of comparison, because they integrate the prior effect of constraints?

2. I would like to fit the same models to multiple datasets (trees or alignments). If I am interested in one global comparison, can I calculate the marginal likelihoods separately per dataset and then simply sum them (as logL) under each model? - Rather than having to fit one hierarchical model to all datasets at the same time.

Thanks in advance for your thoughts!

Best – Jan