3 thoughts on “Do we need an integrated Bayesian/likelihood inference?

  1. I already read the article! Robert posted in his blog earlier than you. And I tryed to figure out which part was written by each one. However, since I've never read anything by Rousseau, I couldn't figure out what was written by her. But I found quite easy to identify you and Robert contribution in the article (though I may be totally wrong).

    Distinguished Gelman styles can be found at:
    references to posterior predictive checks; Bayesian not being subjective, references to hierarchical model (p. 10), argument about using the data twice (due it may be a Robert argument as well)… And my guess is that some exclamation marks are due to you as well! (footnoe on page 4, and "!" on page 11).

    Distinguished Robert Styles are:
    Sound defense of Bayesianism and to take Bayesianism seriously, Bayes factor as a way to compare models, section 4 in general, references to improper priors…

    Am I more right than wrong on this?

  2. This is pretty much what I guessed as well. I found it a bit odd to read a paper with AG as a co-author defending Bayes factors as a useful technique.

Comments are closed.