The Causal Models in Partisans' Minds

• In Event: Experimental Research on Partisan Polarisation and Misinformation

Abstract

Do voters reason that others’ issue concerns are outcomes of a partisan or ideological influence, or instead that issue concerns are causal explanations for others' voting preferences? Under the belief that party or ideology determine issue concerns, an individual will infer that others have correlated intensities in their party's issues: intensity in one issue conveys partisan or ideological intensity, which implies a correlated intensity in other issues. In contrast, under the belief that issues are substantially independent motivations and intensities combine additively to influence a discrete choice, then the opposite is true. Since separate issues provide substitute explanations for the voter's choice of candidate, providing evidence of high intensity in one issue should restrict inference about the intensity of other issues, while evidence of low intensity in one issue will raise expectations that concerns about a different issue motivated the vote choice. Using a controlled laboratory experiment, I find substantial heterogeneity in belief models, and that individuals tend to consistently use one type of model over many different contexts. When the types are pooled, the typical individual appears to have partisan-intoxication beliefs, with a weak but significant response to the information treatments. However, this estimate represents an average behavior of two types with strong responses in opposing directions. Using a panel of decisions, I clustered individuals into types based on their respective Bayesian models and estimated issue-specific model parameters. The structural estimation suggests that although complete “explaining-away” is often not reached, on balance, about half of the individuals perceive a world of issue voters rather than a world of intoxicated partisans. Furthermore, even without performance feedback, experience observing others' diverse issue intensities tends to increase reasoning about instrumental voting.

Author

• Marli Dunietz, Florida State University