Thursday, September 18, 2014

SNA All Star Damon Centola


Damon Centola is the Assistant Professor of Behavioral and Policy Sciences at MIT where his work focuses on the relation between social networks and social contagions. Most of his work since 2009 focuses on how social networks affect health related preferences and can guide policies that want to spread a certain type of positive behavior within a community.

To study how people a desired behavior Centola ran an experiment to see how responsive people were to an online fitness program, depending on the network that the program represented. We may have observed that people with certain negative habits tend to interact primarily with one another, limiting their interaction with others who may have better habits. This makes it difficult to enforce positive changes. This concept is applicable to the health related behaviors as well. People who are less healthy tend to display this ‘homophily’ or similarity among their contacts. Could a network of clustered people then create a faster spreading positive behavioral change? Centola recreated this situation on an online community where he paired participants as “buddies” based on their interests and personal characteristics like age, BMI etc. He then placed participants in two networks: one with long ties and those with just large clusters of people. He found participants in clusters responded much more positively than participants with longer ties. Moreover, social reinforcement was observed when having more health buddies participate in a similar program made a participant return frequently to the program. Therefore closely clustered groups can be faster at spreading a certain behavior, especially in the context of disease prevention. Studies on behavioral implications of social networks can better guide our policies to achieve maximum impact in a short time. “Neighborhood effects” cause better spillovers for desirable health behaviors. The picture below depicts the diffusion of habits in a clustered group (right) versus a group with long ties (left). Clustered group shows faster diffusion (represented by blue dots)



Watch this video to see how the public-health networks can be better targeted for spreading preventative habits!





No comments: