“It is perhaps merely an accident of history that skeptics and subjectivists alike strain on the gnat of the prior distribution while swallowing the camel that is the likelihood”

I recently bumped into this 2013 paper by Christian Robert and myself, “‘Not Only Defended But Also Applied’: The Perceived Absurdity of Bayesian Inference,” which begins:

Younger readers of this journal may not be fully aware of the passionate battles over Bayesian inference among statisticians in the last half of the twentieth century. During this period, the missionary zeal of many Bayesians was matched, in the other direction, by a view among some theoreticians that Bayesian methods are absurd—not merely misguided but obviously wrong in principle. Such anti-Bayesianism could hardly be maintained in the present era, given the many recent practical successes of Bayesian methods. But by examining the historical background of these beliefs, we may gain some insight into the statistical debates of today. . . .

The whole article is just great. I love reading my old stuff!

Also we were lucky to get several thoughtful discussions:

“Bayesian Inference: The Rodney Dangerfield of Statistics?” — Steve Stigler

“Bayesian Ideas Reemerged in the 1950s” — Steve Fienberg

“Bayesian Statistics in the Twenty First Century” — Wes Johnson

“Bayesian Methods: Applied? Yes. Philosophical Defense? In Flux” — Deborah Mayo

And our rejoinder, “The Anti-Bayesian Moment and Its Passing.”

Good stuff.

3 thoughts on ““It is perhaps merely an accident of history that skeptics and subjectivists alike strain on the gnat of the prior distribution while swallowing the camel that is the likelihood”

  1. We’re discussing your “Objections to Bayesian Statistics” and the reply papers in my applied stat class this week – this was a nice supplement for me to come across! I’d imagine it must be interesting for you to go back to these pieces you wrote and think about how the field has changed since then. These papers all get back to the point that one shouldn’t worry about the philosophy behind the model so much as the model’s fit/validity/predictive power, but it’s enlightening to hear it from different perspectives.

  2. Thank heaven for computing power! So many of the old objections were rooted in inability to model rooted in inability to conceive of a model rooted in literal inability to compute. I know that’s reads as sort of circular but it isn’t.

  3. Pingback: Measuring the information in an empirical prior « Statistical Modeling, Causal Inference, and Social Science

Leave a Reply

Your email address will not be published. Required fields are marked *