Clustered Networks And Behavioral Change
Damon Centola of MIT has devised an experiment that seems to show that dense social networks — where people’s contacts were more likely to also be connected to each other — lead to more behavior changes.
Jess McNally, Clustered Networks Spread Behavior Change Faster
To do the experiment, he created an internet-based health community and invited people already participating in other online health forums to join. Over 1,500 people signed up to participate, and they were placed anonymously in one of two different kinds of networks: a random network with many distant ties (above left), or a clustered network with many overlapping connections (above right).
Users in both networks had the same number of assigned “health buddies.” They couldn’t contact their buddies directly, but they could see how their buddies rated content on the site, and could receive e-mails informing them of their buddies activities. Centola said he deliberately didn’t pay the volunteers, so they would participate out of legitimate interest in the site’s content.
In six different trials over a period a few weeks, Centola seeded the site with information about an online health forum and tracked people as they signed up and participated.
In the clustered network, 54 percent of the people signed up for the forum, compared to 38 percent in the random network, and almost four times as fast. Not surprisingly, Centola also found the more friends people had that also signed up, the more likely they were to return to the forum to participate.
Consider this is a social business context: simply by introducing social tools that increase the clustering effect of existing social networks, behavioral change will become more likely. To the extent that the social tools amplify other emotions or perspectives — like the sense that time is passing more quickly in settings where users relate progress on goals — these factors can play together to have a very large impact on business operations.
