Partisanship Over Friendship: Networks and Partisanship in Citizen Forecasting

• In Event: Forecasting Political Outcomes with Different Methods

Abstract

Previous research has indicated that citizens' social networks - their connections with friends, family, and acquaintances - can affect their ability to correctly forecast an election. But it is not fully clear which citizens are most responsive to their social context. Using a unique module from the 2018 United States Cooperative Congressional Election Survey, we examine who responds to their social network's partisan composition when assessing which party is likely to win their district, control of the US House, and control of the US Senate. We find that strong partisanship completely masks the influence of social networks, but even weak partisans are significantly less influenced by networks than their partisan predispositions. However, independents are drastically influenced by their networks when assessing who is likely to win. This research has important implications for our understanding of the utility of citizen forecasting, and the relationship between network and partisan influence.

Authors

• Debra Lynn Leiter, University of Missouri-Kansas City

• Jack Lyons Reilly, New College of Florida

• Mary Stegmaier, University of Missouri