Thursday, December 20, 2012

Statistics: the ecological fallacy

Apparently I'm putting statistics on this blog.  Maybe this will turn out to be a modeling blog; it kind of feels like it already.

The Ecological Fallacy is attributing features at the individual level based on statistics and trends at the group level.  It's not always easy to recognize.

Review of Nate Silver's book

Wow, this is a great review - of how Silver overestimates his own expertise in the financial and medical fields, essentially by failing to model the context in which numerical models are originated and used.