Comment: Bayesian Inference: The Rodney Dangerfield of Statistics?

摘要

All statisticians agree that Bayes' Theorem is the ideal way to approach statistical inference if those involved can reach agreement on the prior and likelihood. Most statisticians are open to some degree of compromise in other cases. Andy Gelman and Christian Robert raise an interesting question: Why then for so many years was Bayesian inference the Rodney Dangerfield of statistics? Why did it get so little respect, including from top intellects like William Feller? But first, was Feller representative ' of more than one idiosyncratic individual?