Conventional wisdom holds that targeting the best-connected individuals in a social network is an effective way to nudge a wider group of people to change their behavior. For example, public-health officials launching a campaign to improve nutrition might target a community’s leaders on the assumption that they wield the most influence. The effectiveness of this approach, however, depends on the structure of the social network, a new Yale study finds. In the study, researchers provide evidence of a structural paradox: When social networks concentrate around a small number of well-connected community members, focusing on those individuals alone does not effectively influence the others’ behavior.

Efforts to change behaviors should look beyond simply targeting leaders to consider the structure of the social circles around the leader.

Yuan Hsiao

“People look to leaders to spread change. But we show that change is not dependent on leaders alone, it also depends on the structure of the social circles around them,” said Yuan Hsiao, assistant professor of sociology and statistics and data science (secondary) in Yale’s Faculty of Arts and Sciences and the study’s lead author. “It matters how people connected to leaders connect to each other. When network ties are concentrated around a few well-connected leaders, it undermines the leaders’ ability to effect large-scale change in behaviors, especially those behaviors that require social reinforcement from many people.”And since individuals in these factions are so drawn to leaders, there is little interconnection among them, he added. “These social barriers restrict the diffusion of new information and behaviors,” he said. “In contrast, we find that behaviors spread better if ties in the network structure are less concentrated around highly connected actors.” The study is published in the American Sociological Review, one of sociology’s premier peer-reviewed journals. Nicholas Christakis, Sterling Professor of Social and Natural Science in Yale’s Faculty of Arts and Sciences, is the study’s coauthor. The study has three parts. In the first, the researchers revisited a classic 1999 study that showed that focusing on well-connected people in a network influences behavior more effectively than targeting random individuals. But this time the researchers adjusted the network structure that study rested upon, which generated different outcomes. Specifically, they showed that the more network ties are concentrated around a few individuals, the less new behaviors spread across the whole network. For the second component, the researchers analyzed empirical data from a 2015 study, which used a randomized trial to test how to enhance adoption of two public health interventions in rural villages in Honduras. (One urged people to use chlorine to purify their drinking water and the other encouraged them take multivitamins to treat nutrient deficiencies.) For each intervention, the authors of the 2015 study examined the effectiveness of targeting well-connected individuals versus random targeting in nine villages. In an analysis, Hsiao and Christakis found that the data from that study supports the hypothesis that when social ties are concentrated around a few individuals, adoption of the new behaviors across the networks decreases.For the third component of the new study, the researchers used computer simulations to explore why the paradox occurs. This revealed that when the social circles of highly connected people are not connected to each other, it reduces reinforcement of new behaviors throughout the social network.“This work has a range of real-word applications,” said Hsiao, who is also a faculty affiliate at Yale’s Institution for Social and Policy Studies. “It can affect how companies market their products, how nonprofit organizations strategize advocacy campaigns, or how public officials initiate policy interventions. “In short, efforts to change behaviors should look beyond simply targeting leaders to consider the structure of the social circles around the leader, which ultimately determines their influence.”