J. Polit. http://doi.org/bznh (2017)
A focus of recent news has been the news apparatus itself, due to concerns over the role of fake news in the US presidential election and questions about the effects of social media ‘echo chambers’. Changing patterns of news production, dissemination and consumption with the advent of social media have also been the focus of recent research, which has highlighted the influence of social network composition on information exposure.
Samara Klar and Yotam Shmargad, at the University of Arizona, carried out a controlled, randomized experiment that addresses confounds inherent in large-scale studies of pre-existing social network data and extends this line of research to preference formation. Participants were assigned to a highly or weakly clustered social network and presented with an opinion piece about a contemporary issue, which they could share with their network. Opinions were seeded at different rates, producing a dominant and an underrepresented view. They found that information diffusion in weakly clustered networks resulted in equal exposure to both views, facilitated learning about the underrepresented view, and led to moderate opinions relative to the highly clustered network, which became increasingly polarized.
Although this is a small-scale study of an artificial network, it points to an important role for network composition in exposure to counter-attitudinal views and preference formation. Understanding these underlying mechanisms will be important in efforts to reduce information bias.