Social Networks – A Superorganism
Nicholas Christakis investigates the role of social networks and their effects in our lives. In his 2010 TED Talk in Long Beach, California, he discusses how we are all embedded in vast social networks, and how one’s location in the network may have a strong, unknown influence on one’s life.
A decade and a half ago at the University of Chicago, Christakis was studying the widower effect: the idea known as “dying of a broken heart.” During his studies, Christakis came across two very apparent details of this effect. The widowhood effect is not restricted only to husbands and wives, nor is it restricted to pairs of people. At this point, Christakis shifted his work to study our embeddedness – number of common neighbors between two nodes – in social networks, and the inherent effect on our lives.
In collaboration with James Fowler, Christakis began to study the epidemic of obesity and its ability to spread from person to person. They constructed network models of dots and edges to represent the social network of 2,200 people in the year 2000. The network is weighted not by edge sign – positive or negative – or strength, but rather by assigning node size proportional to people’s body size. Also, a person’s node is colored yellow if their BMI is greater than 30, or they are clinically obese. At first glance of the network, one can pick out clusters of obese and non-obese people; the clustering coefficient of obese people appears to be high. Yet there is still much that can’t easily be seen and mathematical models reveal the following statistics:
-If one’s friend is obese, his own risk of obesity is 45% higher than it would be otherwise.
-If one’s friend of a friend is obese, his own risk of obesity is 25% higher.
-If one’s friend of a friend of a friend is obese, his own risk of obesity is still 10% higher.
It is apparent the data is a product of the clustering in the network, but what is the cause? Three possibilities to consider are induction, homophily, and confounding. Induction being the spread from person to person, homophily is the idea of obese people choosing to associate with one another, and confounding considers a common exposure to something. Studying the data, Christakis found evidence for all three possibilities. Of course, conducting this study over time could shed more light on the hidden causes of the structure, especially because obesity is a multicentric epidemic – there’s no initial starting node. Most social networks represent a snapshot at an instance, but of course are subject to change over time. An animation of the network over a 30 year period shows constant fluctuations in the makeup of the network.
What is brought to light is an ultimate conclusion of social networks in general: The network has a memory and flows, yet never dies. The network has a resilience that allows it to persist through time. It is important to view social networks as living organisms, or superorganisms, because their properties can’t be understood by merely studying the individuals. The potential for understanding social networks reaches beyond the topic of health but also to the scope of crime, the economy, and warfare.
http://www.ted.com/talks/lang/eng/nicholas_christakis_the_hidden_influence_of_social_networks.html
[…] increases an individual’s chance of becoming obese. This study (also mentioned in another blog) begs the question of how social interactions and relationships influence an individual’s Body […]