5 ms·
Ok, I'm going to assume that this article draws from the work of Christakis and Fowler on social networks and happiness, obesity and depression. http://citeseer
by disgruntledphd2 15y ago
Ok, I'm going to assume that this article draws from the work of Christakis and Fowler on social networks and happiness, obesity and depression. http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.165.3862&rep=rep1&type=pdf http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.165...
This is really interesting work (and i would love it if it were true) but there are some concerns about the statistical methods used in these claims. See Lyons - the spread of evidence poor medicine: http://arxiv.org/pdf/1007.2876 http://arxiv.org/pdf/1007.2876
To sum up, this idea that you can affect behaviour and outcomes based on your friend group is very controversial.
There are two possible explanations: homophily - you surround yourself with people similar to you, and contagion - your opinions affect your peer group. The biggest problem is that these explanations cannot be distinguished on the basis of the kind of studies (retrospective observational) reported by Christakis and Fowler. To do this, you would need a prospective controlled experiment tracking friendships and outcomes over a number of years.
That being said, I would like to believe it, but the evidence is not strong enough for this to be accepted uncritically.
- abhaga 15y agoThis. But while in friend circles, homophily can be a strong factor, the sibling and spouse relationship measurements are interesting since homophily cannot affect those to the same extent.
- disgruntledphd2 15y agoSibling I'll give you, but there's a lot of evidence to suggest that people choose spouses who are more similar to themselves, so I'm not sure how much of a role contagion (if its real) can play in this effect.
- sskates 15y agoIt's a difficult effect to do a controlled experiment for because of the way homophily acts as a confounding variable. Anecdotal evidence suggests the effect is real though- as Justin said in the article, he didn't actually start riding a motorcycle until he met other people who did.
- disgruntledphd2 15y agoIts doable though. You recruit over a number of years at a college, get them to indicate who their friends are and ask them to fill out some personality questionnaires and demographics. Follow them for their college career, repeating the measures and asking them for their closest friends at each time point (say every month or semester). It would be expensive, but its very achievable. Unfortunately, anecdotal evidence is not particularly useful when dealing with people, given that we tend to reinterpret our pasts based on our presents among other memory biases.