How does network structure influence opinion? Relying on theories of preference formation and social networks, we randomize a sample of adults into networks that vary in structure. In one (a clustered lattice), individuals' connections tend to be connected to each other; in another (a random network), individuals' connections tend not to be connected, instead providing access to different regions of the network. We seed messages that reflect competing sides of policy debates in each network: one underdog viewpoint is seeded less often, while a dominant viewpoint is seeded more often. We track their diffusion and find that the random network increases exposure to underdog views, compared with the clustered lattice. Individuals in the random network subsequently learn more about the policy debates and become more sympathetic toward the underdog perspective. This has implications for how less funded information campaigns can strategically target social networks to maximize exposure and change minds.
ASJC Scopus subject areas
- Sociology and Political Science