A caricature of Bayesian statistics (2)
The letter to the editor of La Recherche I posted there has just been published in the November issue (No. 424, p.6). It also includes a reply from Nicolas Vayatis who, while protesting he did not want to present Bayesians as sectarians, maintains his anecdote about the Bayesian songs sung at the beginning of each Bayesian conference! He also argues that the difference between Bayesians and frequentists is primarily philosophical and sociological (?!). The conclusion of the reply is somewhat tongue-in-cheek, in that Nicolas Vayatis states that his main message for young statisticians is to keep in touch with societal issues by developping modelling tools. This still sounds critical of communitarian (or even sectarian) tendencies, while the rapid development of Bayesian statistics in the past twenty years is exactly due to the involvement of Bayesian statisticians in all directions of statistical modelling. Anyway, there is no need to keep the ball rolling now that the point has been made.