What is Bayesianism? Why should you care?

I’m hugely grateful to Ignite Bristol for allowing me to open their second night with this 5 minute talk about probability, and to the film crew for doing such a professional job.

Though lots has been written about Bayes, I wanted to convey to a lay audience what he achieved and why it’s so important now. Here is an attempt at a set of “footnotes” for anyone who wants to follow up:

The historical background is explained in:

  • John Earman, Bayes or Bust (MIT Press)
  • Edwin T. Jaynes,  Probability Theory: the logic of science (Cambridge University Press) and his other writings.

The Jaynes book suggests some of the applications of Bayesianism that I mention, including radio astronomy. It was my own idea to include flirting as an example of a reverse inference problem: I’m sure people don’t usually think of it that way.

“Is Bayesianism just common sense…?” For examples of how medical or other decisions go wrong because people don’t think in a Bayesian way, an accessible introduction is Stuart Sutherland, Irrationality (Penguin paperback, reissued by Pinter & Martin).  Similar material is covered in Chapter 12 of Scott Plous’ The Psychology of Judgement and Decision Making (McGraw-Hill). These authors in turn, are digesting the work of Amos Tversky, Daniel Kahneman and colleagues.

A full-on technical overview of Bayes’ theorem, why it should be taken seriously as a standard of reasoning, and how human judgements violate it, can be found in Jonathan Baron’s textbook Thinking and Deciding (Cambridge University Press).

According to my memory, although I can’t find the reference, the “Bayesian argument against torture” was proposed by the mathematician Robert Matthews.

“The training had absolutely no effect”: This remark conflates a number of different studies; I had to because of the limitations I was speaking under. Two papers reprinted in the book Judgment under uncertainty: Heuristics and biases describe how attempts to train people to be more Bayesian often made no difference: Alpert & Raiffa’s “A progress report on the training of probability assessors” and Lichtenstein, Fischhoff & Phillips’ “Calibration of Probabilities: The State of the Art.”

, , , , , ,

  1. #1 by NoumenalRealm on November 15, 2010 - 10:04 pm

    Thanks for this list of intro books. As a lay but eager to learn, person. I’ll keep this in mind. I’ve often been in a few talks on Bayesian-related issues without having a clue about all the symbols.


Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: