Philosophie et practique de la statistique bayésienne. I’ll try to update the slides a bit since a few years ago, to add some thoughts I’ve had recently about problems with noninformative priors, even in simple settings.

The location of the talk will not be convenient for most of you, but anyone who comes to the trouble of showing up will have the opportunity to laugh at my accent.

P.S. For those of you who are interested in the topic but can’t make it to the talk, I recommend these two papers on my non-inductive Bayesian philosophy:

[2013] Philosophy and the practice of Bayesian statistics (with discussion). {\em British Journal of Mathematical and Statistical Psychology} {\bf 66}, 8–18. (Andrew Gelman and Cosma Shalizi)

[2013] Rejoinder to discussion. (Andrew Gelman and Cosma Shalizi)

and this paper on Bayesian attitudes:

[2008] Objections to Bayesian statistics (with discussion). {\em Bayesian Analysis} {\bf 3}, 445–450. (Andrew Gelman)

[2008] Rejoinder to discussion. {\em Bayesian Analysis} {\bf 3}, 467–478. (Andrew Gelman)

and also these two recent papers on the importance of informative models in routine Bayesian inference:

[2012] P-values and statistical practice. {\em Epidemiology}. (Andrew Gelman)

I liked the 2014 paper. The Bayesian approach to variability (or almost anything really) is a sound one, especially given sparse strata.

But the distinction btw modeling the variability, and explaining it, not so clear. What causes the heterogeneity? In my view this is not an estimation problem.

In Cochran’s last paper, the last line is something like when “effects are different in

different places and at different times” it is very puzzling.

I believe he struggled ever since Yates and he wrote a 1938 paper on repeated agricultural studies.

Shame that issue got buried so much for so long.

[…] reader calls my attention to Andrew Gelman’s blog announcing a talk that he’s giving today in French: “Philosophie et practique de la […]