Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

miro comments on Useful Statistical Biases - Less Wrong

11 Post author: Eliezer_Yudkowsky 01 April 2007 04:51AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (4)

Sort By: Old

You are viewing a single comment's thread.

Comment author: miro 23 April 2009 11:51:56AM 0 points [-]

I have a somewhat related question - and openly admit to being a neophyte

my question is this traditional variance weights positive and negative outcomes equally

how can one compute a variance that reflects a persons bias (risk aversion) toward a directional outcome as in business assume an ill favored outcome is worth 0.5x and a preferred outcome is worth 1.5x

would a person compute 2 variances by creating 2 sub populations illfavored/preferred and apply the formula var (bx) = b^2times sigma^2 to each population and sum the final products?

am I wrong in this line of thinking? is there another approach? its been quite some time since my university stats days - so please be gentle with my ignorance

appreciate your thoughts and if you ping my email to let me know miroslodki (at) yahoo (dot) ca

btw - fascinating site and discussion regarding crowd wisdom - fwiw I share your viewpoints/concerns you've found a new reader cheers Miro