Bayesian Truth Serum

nel.mit.edu

39 points·compbio··5 comments

It looks like an algorithmic leveraging of the Dunning-Kruger effect.

Yes, it looks similar.

Problem: Getting honest questions on questionaires is difficult.

Setup: Respondents supply not only their own answers, but also percentage estimates of others' answers.

Algorithm: The BTS formula assigns higher scores to answers that are surprisingly common.

Premise: If people truly hold a particular belief, they are more likely to think that others agree or have had similar experiences.

http://www.slideshare.net/micmickimo/introduction-to-bayesia...

I don't see why that would be. It seems to me that people are vastly more likely to purport to hold an orthodox opinion when they actually hole a heretical one than vice versa. But obviously they will also predict that other people will purport to hold the orthodox opinion.

That should only be the case when they are poor analysts. If they are good analysts, they should be more likely to claim a heretical opinion and predict orthodox opinions.

After all, those who have the most information are not in a position to accept the most common opinions, and they have enough information to know what the common opinions are.

I would say it's not really about who is telling the truth, but who you should listen to -- who is relevant in addition to truthful.

I think gweinberg had in mind the sort of situation in which saying you hold a heretical opinion brings a punishment that would not be offset by you being recognized as a superior expert later. This creates an incentive to lie.

Reply on news.ycombinator.com