Tuesday, February 03, 2009

Genes and Social Networks   posted by Piccolino @ 2/03/2009 04:13:00 PM

Heritability estimates are slippery animals, but this recent PNAS paper is a great illustration of how they can be used to discipline theories of social network formation. The authors start by showing that three building blocks of social networks are heritable, namely the number of friends you have, the number of people who name you as a friend, and the likelihood that two of your friends are also friends. They then ask if existing theories of social network formation are consistent with empirical fact that a large share of individual variation in these buildling blocks is explained by individual characteristics. Perhaps not too surprisingly to readers of this blog, a model which allows individuals to differ ex ante does considerably better than models which make a blank slate assumption. The paper also fits in nicely in the tradition of behavior genetic work which emphasizes how people based on their inherited traits self-select into particular environments. Razib pointed out the other day that we can taxonomize traits into those whose genetic architecture we understand pretty well (skin coloration) and those that are still a puzzla (IQ). I am curious to see where the social network building blocks fit in. I have no doubt we will have the answer to this question very soon.