Search results

Abstract
A popular explanation for America's democratic ills is that Republicans and Democrats misperceive one another to hold extreme attitudes. However, Americans may also misperceive the diversity of partisans' attitudes to ill effect. This paper uses surveys and pre-registered experiments with representative and convenience samples (n = 9405) to validate a measure of perceived attitude extremity and diversity and compare it to canonical measures of perceived polarization. We find that American partisans vastly under-estimate the diversity of each party's attitudes. Yet, contrary to existing research, we see little evidence that partisans over-estimate how extreme the "average" Republican or Democrat is. Finally, perceptions of both the "average" partisan and within-party attitude diversity predict partisan animosity and perceptions of out-party threat.

Abstract
Many warn that the United States is on the brink of democratic collapse, because partisan animosity, support for partisan violence, and support for undemocratic practices are on the rise. Quelling some fears, scholars have offered interventions that use messages to correct misperceptions about citizens' partisan opponents (the "out-party"). In this article, we provide evidence that the effects of these interventions are not as robust or consistent as hoped. First, we use panel data (n=9,810) to show that perceptions of the out-party are highly variable. This suggests that these perceptions are weakly held and thus unlikely to be a significant cause of hostile attitudes. The oscillation of perceptions over time also suggests that, for many, any effect of corrections would likely be overwhelmed in just 1 month. Second, in a meta-analysis of 67 statistical tests from 12 studies in eight papers, we document that current evidence on the efficacy of corrections is weak. Third and finally, in pre-registered experiments (n=2,846), we find that changing Americans' perceptions of the out-party's demographics, policy attitudes, and support for undemocratic practices has no consistent effect on partisan animosity, support for partisan violence, or support for undemocratic practices. These observations suggest that correcting misperceptions of the out-party is not a panacea for our democratic ills.

Political interest is a key predictor of likelihood to vote. We argue that the political interest–vote intention relationship can be explained by well-established theories that predict behavior across domains (e.g., theory of reasoned action, theory of planned behavior). Relying on the integrated behavioral model, we propose a core mediation model with vote likelihood (i.e., behavioral intention) as the dependent variable. Two types of media use (conservative and liberal–moderate) are then assessed in relation to the core model. We explore the ways in which our results contribute to theory and outline a research agenda.