The Spread of Obesity in a Large Social Network
Link to original article:
http://www.nejm.org/doi/full/10.1056/NEJMsa066082#t=article
While this article on the spread of obesity is relevant in the sense that it shows how network effects propagate from local to global, it also portrays how biological, social, psychological and geographic relationships and ties are powerful harbingers of change. I will mainly discuss social ties here.
Introduction
The study was done on a network of 12,067 people who underwent repeated measurements over a period of 32 years. Various factors were examined several aspects of the spread of obesity, including the “existence of clusters of obese persons within the network, the association between one person’s weight gain and weight gain among his or her social contacts, the dependence of this association on the nature of social ties and … and geographic distance between the domiciles of persons in the social network. ”
Social influences
It might come as a surprise that something like obesity can “spread”. What “spread” here means is the ability of the subject (called “ego”) to influence his/her connections (called “alter”). This “spread” often results in a cluster of obese people, thus spreading from one person to thousands of people. The three main reasons the article discusses for this “spread” are
- Homophily: egos might choose to associate with like alters
- Confounding: egos and alters might share attributes or jointly experience unobserved contemporaneous events that cause their weight to vary at the same time
- Induction: alters might exert social influence or peer effects on egos
This is very easy to associate with; how many of us have taken classes or gone for events because our friends are going? That this pulling or pushing of relationships can be extended to obesity is something that is the marvel of networks!
I mention some more interesting facts here: firstly, it seems that a mutual friendship between the ego alter has the most influence, nearly a 200% increase! It signifies this graph has directed edges, and the influence of A on B is not the same as that of B on A. This ties (pun intended) back to the fact that a stronger mutually beneficial relationship is likely to result in more influence for both sides.
Figure 3B indicates that the effect of geographic distance is different from that of social distance. Whereas increasing social distance appeared to decrease the effect of an alter on an ego, increasing geographic distance did not have any changes.
“These results suggest that social distance plays a stronger role than geographic distance in the spread of behaviors or norms associated with obesity.”
Concluding remarks
This experiment does an excellent job at showing how networks are extensively used in real life for modelling. This research has been done extremely carefully accounting for statistical errors. Using technology to graph this network makes it visually appealing. I attach one of the largest connected component:
While we may discuss how powerful this study is, we must make sure we are doing something to counter the risks it enlists. Obesity is one of the world’s largest killers; given massive data and so much insight into the network’s tangled mess, it would be really interesting to find ways to stop such “domino” effects from spreading by cutting at the point of contact, a kind of “inverse network” strategy.
This research article thus has a mix of the following important network concepts covered in class:
local to global, strength of ties, directed edges, social influences, geographic distances
I really hope all readers of this blog post do take time to read the actual report. It truly is very enlightening. 🙂