peer-reviewed articles

2025

1. Unraveling a “cancel culture” dynamic: When, why, and which Americans sanction offensive speech

   Nicholas C Dias, James N Druckman, and Matthew S Levendusky

   Journal of Politics, Apr 2025

   Abs DOI

   There is a growing belief that many Americans shun, ostracize, or “cancel” those they dislike or who make disagreeable statements. Yet, no empirical work has explored the prevalence or motives of this type of sanctioning or how Americans perceive it. Using a nationally representative survey with an embedded conjoint experiment, we find that Americans vastly overestimate how likely other people—especially out-partisans—are to cancel others. Nevertheless, they accurately perceive what motivates others to cancel: disagreeable and offensive statements, not disliked speakers. Additionally, we find that Democrats and Republicans are similarly motivated to cancel, though canceling behavior out in the world may more commonly come from Democrats. Our findings highlight how “cancel culture” could limit harmful speech, but encourage self-censorship and partisan animus. They also reveal the normative fault lines underlying debates about free speech in contemporary society.

2024

1. American partisans vastly under-estimate the diversity of other partisans’ policy attitudes

   Nicholas C Dias, Yphtach Lelkes, and Jacob Pearl

   Political Science Research and Methods, Sep 2024

   Abs DOI

   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 = 9,405) 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.

2. Correcting misperceptions of partisan opponents is not effective at treating democratic ills

   Nicholas C Dias, Laurits F Aarslew, Kristian Vrede Skaaning Frederiksen, and 3 more authors

   PNAS Nexus, Aug 2024

   Abs DOI

   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.

2022

1. The nature of affective polarization: Disentangling policy disagreement from partisan identity

   Nicholas C Dias, and Yphtach Lelkes

   American Journal of Political Science, Jul 2022

   Abs DOI

   Democrats and Republicans clearly dislike one another. Yet, scholars debate whether policy disagreement or partisan identity, per se, drives interparty animus. Past studies suggest the relationship between partisanship and interpersonal affect is spurious, driven by inferred policy preferences. We argue, instead, that policy preferences signal partisan identity when the parties’ stances on an issue are well-known. Using a nationally representative survey and four preregistered experiments, we disentangle the effects of policy disagreement and partisan identity on interpersonal affect. Our findings suggest that partisan identity is the principal mechanism of affective polarization, and that policy preferences factor into affective polarization largely by signaling partisan identity. However, our results also affirm that policy disagreement in itself drives interpersonal affect. This provides evidence that partisanship reflects an emotional attachment to a political party, not merely a running tally of rational considerations.

2021

1. Exploring the role of media use within an integrated behavioral model (IBM) approach to vote likelihood

   R Lance Holbert, Nicholas C Dias, Bruce W Hardy, and 6 more authors

   American Behavioral Scientist, Mar 2021

   Abs DOI

   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.

2020

1. Researching fact checking: Present limitations and future opportunities

   Nicholas C Dias, and Amy Sippitt

   Political Quarterly, Jul 2020

   Abs DOI

   The fact checking field has grown tremendously in the past decade, as has academic interest in the practice, with dozens of studies testing the effectiveness of corrections. However, research on fact checking is not yet optimised to help fact checkers address the global challenges of mis- and disinformation. In this paper, we review the literature on fact checking’s effects and identify two key gaps. First, we discuss the limited diversity and external validity of existing studies, which have overwhelmingly been conducted in Western countries and under artificial, experimental conditions. Second, we argue that research has narrowly focussed on the short-term, corrective effects of individual fact checks, largely ignoring the multiple ways fact checkers conceive of their impact. Thus, research has overlooked the cultural and systemic changes that fact checkers pursue. We conclude by highlighting opportunities for further research and for improving communication between academics and fact checkers.

2. Emphasizing publishers does not effectively reduce susceptibility to misinformation on social media

   Nicholas C Dias, Gordon Pennycook, and David G Rand

   Harvard Kennedy School Misinformation Review, Jan 2020

   Abs DOI

   Survey experiments with nearly 7,000 Americans suggest that increasing the visibility of publishers is an ineffective, and perhaps even counterproductive, way to address misinformation on social media. Our findings underscore the importance of social media platforms and civil society organizations evaluating interventions experimentally rather than implementing them based on intuitive appeal.

working papers (available upon request)

2025

1. Do inequalities in factual policy knowledge matter?

   Nicholas C Dias

   2025

   Abs

   Public opinion is meant to reflect all citizens’ values and interests. But some argue that public opinion is not what it seems: Some citizens have more facts than others, enabling them to better translate their abstract principles into concrete policy opinions. Consequently, public opinion over-represents the wishes of these informed citizens. Yet, existing measures of knowledge are incomplete and biased, suggesting that knowledge inequalities may arise because tests focus on facts that are more relevant to certain groups. To overcome these limitations, I propose a novel approach to measuring how citizens understand policies, which examines whether citizens recognize different rationales for policy opinions. Then, in a demographically representative sample of 2,000 U.S. adults, I use this measure to reexamine whether some groups are better equipped to form policy opinions. Contrary to past evidence, I find that citizens with different education, political interest, and values understand policy issues in similar terms.

2. Are value-consistent opinions more stable over time? Not really

   Nicholas C Dias

   2025

   Abs

   Since Converse’s classic work, scholars have assumed that temporally unstable opinions are non-attitudes: shallow opinions that poorly reflect a citizen’s core values or interests. But this assumption has not been carefully scrutinized. This article offers a broad test of whether “value-consistent” opinions—i.e., opinions that do align with a citizen’s values or interests—are more stable over time. Within high-quality panel datasets, I identify pairs of values and policy issues wherein the value entails an opinion about the issue. I then examine whether citizens are more likely to hold onto policy opinions that align with their values. I find that they consistently do—even after controlling for panelist, policy issue, value, and period fixed effects. However, the size of this effect is small, suggesting that temporal stability is a weak signal of how value-consistent an opinion is. As such, scholars should rethink the implications of the vast literature on opinion stability.

3. Can citizens use information to form principled opinions?

   Nicholas C Dias

   2025

   Abs

   Many scholars argue that citizens’ policy opinions are a poor reflection of their underlying values and interests. This claim is troubling, as it implies citizens may vote against their principles. Yet, existing tests of are likely to underestimate the extent to which citizens form principled opinions. First, these methods cannot account for principles that the researchers did not anticipate. Second, citizens are often asked to evaluate policies whose relation to their principles is undefined. I address these problems with a new method, which examines whether citizens use information about policies’ consequences to align their policy opinions and principles. Then, I demonstrate this approach with a pre-registered experiment, which centers on citizens’ values of distributive justice. I show that citizens broadly prefer policies that distribute benefits according to their distributive values. As such, this article calls on researchers to reexamine their pessimism about whether citizens’ opinions can meaningfully guide government policy.