Social Network Analysis: Interesting But Still Limited

“Connected” by Christakis and Fowler explains the benefits of social network analysis with fascinating and important stories. Social network analysis tells us who is most likely to catch a habit. This is just like how doctors can predict who is most likely to catch a disease from the person’s place in a network. The analysis of networks helps illuminate a vast range of fields. Marketers should be interested, as well as medical doctors, political campaigners and high school principals.

They offer general observations such as the Three Degrees of Influence Rule. Which means we influence our friends, we also influence our friends’ friends (degree two), and their friends (degree three). Our influence tends to fade to such a degree that this friend is the limit of our impact.

The book also shows great visualizations of networks. These explain how to interpret networks and the effects that network structures have on the network’s members.

That said social network analysis has limits. It finds it hard to provide the sort of satisfying theoretical explanations that we really should have to truly understand what is happening. For example, Christakis and Fowler tackle the problem of voting, namely why does anyone bother to vote when the chance of impacting the election’s outcome is infinitesimally small. Their idea is to use social network analysis to get away from explanations like voters feel satisfaction from doing civic duty etc… They suggest a reason why people vote is that “they are connected” (Christakis and Fowler, page 192). Our actions influence others so “Instead of each of us having only one vote, we effectively have several” (Christakis and Fowler, page 191). We influence other people, who tend to be like us, to vote. We are “therefore much more likely to have an influence on the outcome of an election” (Christakis and Fowler, page 192). The theory here is quite weak. One vote out of ten million has almost no influence but even if we “control” five votes this is still incredible unlikely to make a difference to an election, five out of ten million is nothing. Studying connections doesn’t really solve the problem of why people bother to vote.

Social network analysis is interesting but not magical — it can’t solve all the problems that researchers have but it can give interesting insights.

Read: Nicolas A. Christakis and James H, Fowler (2011) Connected: How Your Friends’ Friends’ Friends Affect Everything Your Feel, Think, And Do ,Little Brown and Company