|
||
Title: Bayesian Truth Serum Post by amichail on Oct 20th, 2006, 9:00pm What do you think of this? http://targetyournews.com/?cmd=surf&urlid=656255 |
||
Title: Re: Bayesian Truth Serum Post by Whiskey Tango Foxtrot on Oct 20th, 2006, 9:15pm If I am reading this correctly, he is using truth serum to establish a basis for evaluating a subjective opinion based on pre-established statistical expectations. It does not seem too awe-inspiring, but maybe I'm hard to impress. Or just misinterpreting it. :-/ Here is a direct link to the article if you can't find it. http://cipd.mit.edu/documents/publication_files/prelec2004science.pdf |
||
Title: Re: Bayesian Truth Serum Post by amichail on Oct 21st, 2006, 4:42am The task of filling out surveys has been turned into a game where participants try to obtain the highest score possible. The best strategy for winning this game is to tell the truth as best as you can. The whole approach is based on a psychological observation that if you prefer X, then you would overestimate the percentage of people who would also prefer X. The specifics are a bit unclear to me though. Maybe someone could elaborate on why this works, at least in the case where questions have two answers. |
||
Title: Re: Bayesian Truth Serum Post by towr on Oct 21st, 2006, 1:07pm on 10/21/06 at 04:42:32, amichail wrote:
I usually assume no one else likes the stuff I like.. But then I've never been that good at playing psychological games ;) |
||
Title: Re: Bayesian Truth Serum Post by Grimbal on Oct 22nd, 2006, 6:36am on 10/21/06 at 13:07:12, towr wrote:
I thought I was the only one like that! ;D |
||
Powered by YaBB 1 Gold - SP 1.4! Forum software copyright © 2000-2004 Yet another Bulletin Board |