Social Networks Based on Genes

Regina George, Serena Van Der Woodsen, and Marissa Cooper may have more to thank for their popularity than just great hair. A group of researchers from Harvard and the University of California, San Diego have suggested that social network structures have a genetic basis—meaning that popularity may be coded in one’s

DNA.

According to co-author and Harvard Professor of Medical Sociology Nicholas A. Christakis, the findings expand on the common intuition that genes influence social behavior, accounting for the variability in an individual’s popularity.

“The interesting point is not that the number of friends varies from person to person,” Chrisakis said, “[rather] that other higher-order aspects of network structure also appears to be genetically determined.”

These high-order aspects of social networks include transitivity, or how well one’s friends know each other, and centrality, the tendency for an individual to be near the center verses the periphery of a network.

Using data on the social networks of 1,110 adolescent twins, researchers were able to show that social networks of identical twins share more similarities in structure than those of fraternal twins in measurements of transitivity, centrality, and the number of friends.

According to Christakis, there are distinct evolutionary advantages and disadvantages of being at a particular location in a social network. He said that if a deadly germ is moving through the population, being in the periphery has a clear advantage. Alternately, being in the center of a social network can provide access to valuable information about prey.

The research questions the validity of current network models that treat individuals as interchangeable nodes and neglect the genetic influence on network structures.

“Our findings suggested that it’s not okay to treat people as interchangeable nodes. There’s heterogeneity at the level of nodes of network that needs to be accounted for,” Christakis said.

The researchers developed a new model, called the “attract and introduce” model, which outperformed the previous models by including heritable traits of social networks. This new model is based on two parameters: people’s attractiveness level and their tendency to introduce their friends to one other.

“The basic idea is that if people are interconnected, their healths are interconnected, if we really want to understand health, we have to understand society,” Christakis said.

The new findings might also have policy implications, according to Christakis.

“If things spread in networks, if I get you to behave well, others will start to behave well too,” he said. “The dollars spent on getting you to behave well have a much bigger rate of return than I previously thought, for instance.”

The paper was co-authored with James H. Fowler ’92 and Christopher T. Dawes, both researchers at the University of California-San Diego.

—Staff writer Gordon Y. Liao can be reached at liao@fas.harvard.edu.