This is a very interesting article on Wired about networks. Two American social scientists found old notes relating to a long-running heart disease study started in 1948 in the Massachusetts town of Farmingham. What was invaluable about this particular study was that it contained notes about who was friends with whom, and also kept track of births, deaths and family relations.
This is a gold mine for social network researchers, as it allows researchers to match detailed health data with social time-lines. The picture that has emerged should not surprise those interested in network science, but it is remarkable for its accuracy. The data shows that our networks determine health-related lifestyle choices more than we would like to imagine. For example, In 1971, 65% of Farmingham adults smoked regularly. By 2001 only 22% smoked, but what became evident was that people responded to what their network was doing. Data showed clearly that people quit together, or they did not quit at all. Similarly, obesity increased on a social-related fashion. Obese people tend to have obese friends; as it seems like people discovered fast food in groups.
This may sound deterministic, but it seems to bear out with the data.