A Bayesian Take on Julian Assange

fivethirtyeight.blogs.nytimes.com

19 points·azazo··3 comments

This is just a pet peeve of mine: journalists who obviously didn't double check the facts they speak of.

The psychological fallacy that the author speaks of is actually well studied and known as the conjunction fallacy and is not really that related to bayesian inference[1].

It is also notable that human decision making shows base rate neglect (not considering conditional probabilities), but when things are presented in terms of frequencies rather than strict probabilities, human decision making follows a bayesian curve[2], thus showing that humans do in fact incorporate bayesian learning and are in fact good at it contrary to what the journalist says.

[1]http://en.wikipedia.org/wiki/Conjunction_fallacy

[2]A. Tversky & D. Kahneman: "The framing of decisions and the psychology of choice", Science, 211, 453-458 (1981).

Excellent point. I think the article puts on the trappings of science, but is actually entirely speculative.

You have a very well formatted post with excellent footnotes, which makes me inclined to trust it. Yet the author of the piece, Nate Silver, is widely regarded as a professional statistician (with a solid degree in economics) and likely has a strong understanding of these issues. I'm predisposed to trust Silver over an unknown source from the internet because of my prior experience with his writings and the respect which others accord him. Therefore, despite the plausibility of your criticism, I still side with Silver.

p.s. My silly self-referential post aside, I think that neither my assessment nor Silver's is an example of the conjunction fallacy.

Reply on news.ycombinator.com