Judges may express underlying political preferences when speaking that can only be captured with audio and not text. Yet, it is unclear if audio or video recordings of judicial proceedings shape high courts' legitimacy differently than written transcripts. We address this question with two survey experiments using short, real-world case excerpts from the US and UK. In sum, we do not find evidence that the method of delivery is associated with evaluations of court legitimacy as predicted by past studies. These null results persist when we account for the emotions that judges convey, perceived political inclinations of judges, and participants' political preferences. Our findings offer novel insight into the relationship between high court attitudes and the medium through which judicial decisions are communicated.
We investigate whether large language models (LLMs) threaten democracy through their persuasive capabilities. Using two survey experiments (N = 10,417) and real-world simulations, we compare the cost-effectiveness of LLM chatbots against traditional campaign tactics, taking into account both the "receive" and "accept" steps in the persuasion process. Our design advances prior research by assessing extended human-LLM interactions and measuring short- and long-term effects across three political domains. We find that while LLMs are comparably persuasive to campaign ads once seen, real-world impact depends on both message reception and acceptance. Simulations estimate LLM-based persuasion costs $48-$75 per voter versus $100 for traditional methods. However, traditional methods currently scale more effectively. While LLMs do not yet offer substantially greater potential for large-scale persuasion, this may shift as capabilities improve and techniques for scalable exposure become feasible.
Do exclusionary signals push ethnic minorities out of politics or mobilize them to politically act together? We study this question with a novel survey experiment among Muslim Turks in Germany that randomized videos of anti-Muslim hate-crime reports (social exclusion) and of an anti-Muslim far-right party's electoral gains (political exclusion). We find that both treatments increase intended political participation, heighten in-group solidarity, and shift vote intentions toward left-wing parties. Crucially, these effects are concentrated among highly integrated Muslims, the very group that prevailing theories of integration predict to be least inclined toward ethnic voting. This pattern suggests that integration is not only a matter of individual resources or time in the host country; it also depends on how the majority treats minority communities. These findings highlight the role of public rejection signals in shaping political behavior, underscoring that even well-integrated citizens may mobilize collectively and support parties they perceive as more protective.
We examine whether exposure to government assistance can generate positive policy feedback effects among constituencies not traditionally supportive of the welfare state. Focusing on the U.S. Paycheck Protection Program (PPP), a small business bailout enacted during the COVID-19 pandemic, we study how making government assistance salient affects attitudes toward social assistance programs among a typically Republican-leaning and relatively affluent group. Leveraging a bespoke survey of verified program recipients and an embedded experimental manipulation, we find that reminding PPP recipients of program participation increases support for government spending on healthcare, nutritional assistance, and unemployment benefits by an average of 6.9 percentage points-equivalent to roughly 16 percent of the partisan divide on these issues. The findings provide novel causal evidence that making the receipt of government assistance salient increases support for anti-poverty programs among well-off people, even when those programs do not directly benefit them.
How do Americans react to perceptions of racial inequality? We subtly introduce economic inequality in an experiment comparing Black and White groups, while varying whether the inequality occurs by chance or is the result of human agency. Subjects are given the ability to correct that inequality by taking actions that could be costly or not. All subjects are more likely to correct inequality if their racial ingroup is disadvantaged. Black subjects react more strongly when inequality is human-made, whereas the source of inequality does not matter to Whites. When they perceive their ingroup as being treated unfairly, Black subjects are more likely to bear personal costs to correct inequality than White subjects. Subjects' concept of fairness switches depending on whether their ingroup or outgroup is disadvantaged: they become more likely to behave unfairly themselves to correct inequality against others if the outgroup benefits at the expense of the ingroup.
Rural Americans constitute a politically consequential yet theoretically understudied identity group. This study reconceptualizes descriptive representation to include place-based identities and demonstrates its influence on policy support and political trust. Using a preregistered, original survey experiment of rural respondents, we assess whether rural Americans exhibit greater support for laws and perceive it as more beneficial to rural communities when proposed by state representatives who share their rural identity. Our findings strongly support this hypothesis: rural Americans express higher levels of support for laws that were introduced by descriptively representative lawmakers and are more likely to believe such policies benefit rural areas. Moreover, respondents demonstrate higher levels of trust in rural lawmakers even in the absence of additional information about them. These results illustrate that, for rural Americans, place-based identity is deeply influential in shaping their political perceptions.
Do men respond to a masculinity threat by adopting more conservative political attitudes? A highly cited 2013 study by Willer et al. - drawing on substantial work in social psychology - argues in the affirmative, reasoning that endorsing conservative views allows men to reaffirm their gender identity. In two experiments with student convenience samples (N total 100-110, N men 40-51), the authors find consistent evidence: inducing masculinity threat increases support for war, homophobic attitudes, and endorsement of dominance hierarchies. We conduct a preregistered replication of this foundational study using a nationally representative probability sample (N total 2774, N men 2073). Contrary to original findings, we observe no consistent evidence that masculinity threat alters political attitudes. We further do not find support for design differences between the replication and original study driving contrasting findings. Our results call into question the robustness of evidence linking masculinity threat to political attitudes and underscore the importance of re-evaluating widely accepted findings with representative, large samples.
Can survey experiments replicate real-world behavioral intent and behavior? I study a population in rural Bangladesh (N similar to 1600) along the banks of the Jamuna River, at risk of riverbank erosion and flooding. I compare their responses to questions about hypothetical movement behavior in vignette-experimental natural disaster scenarios (pre-monsoon, May-June 2021) with their migration intentions and actual migration 2-6 months later, following quasi-experimental real-world exposure. My results show that hypothetical as well as actual affectedness and risk shape migration intent and behavior in structurally similar ways, indicating sign-generalization over both treatments and outcomes. However, the vignette experiment approximates actual behavioral intent more closely than behavior, suggesting that real-world intention-behavior gaps can complicate external validity. Given a slim evidence base for generalizability over treatments and outcomes, this study contributes a crucial comparison from a rarely studied developing-country context on what we can learn from survey experiments.
Stereotypes about groups are commonly measured by asking participants to rate the groups on a scale. However, the percentage of participants who stereotype a group can be affected by the order in which participants are asked to rate the groups. Data from a randomized experiment in the American National Election Studies 2022 Pilot Study indicated that a group was more frequently positively stereotyped relative to another group when the group was asked about first in the pair of groups, compared to when the other group in the pair was asked about first. Researchers are therefore advised to randomize the order of groups in a stereotype battery to evenly spread this ordering effect across groups and are also advised to design stereotype items to minimize this ordering effect.
How do citizens in Sierra Leone perceive the legitimacy and feasibility of sanctioning their chiefs outside of elections? This study investigates perceptions of non-electoral sanctions through a pre-registered survey experiment in Sierra Leone. We find that citizens view indirect sanctions – such as appealing to higher authorities – as more legitimate and feasible than direct actions, and that the range of acceptable sanctions expands with the severity of the offense. Community elders’ involvement increases the perceived legitimacy of sanctions, highlighting their role as political intermediaries. Finally, respondents’ social status moderates their perceptions of both the legitimacy and the feasibility of sanctions. These results suggest that even in highly hierarchical settings, citizens may retain some capacity to discipline chiefs, though accountability seems primarily mediated through vertical institutions rather than direct collective action.
This study examines the behavioral consequences of partisan group composition on cooperation in a setting where cooperation is mutually beneficial but unethical. Collaborative corruption highlights that corruption is not a solitary act but necessitates cooperation. Based on the premise that partisanship serves as a social identity, leading ordinary citizens to reward co-partisans and penalize out-partisans, we expect that collaborative corruption is higher in partisan-wise homogeneous groups. To test this expectation, we conducted a preregistered, large-scale experiment among U.S. voters playing an online version of the collaborative cheating game by Weisel and Shalvi. We find no evidence that partisan homogeneity affects collaborative cheating. These results suggest a critical scope condition: while partisan homogeneity improves cooperation in social dilemmas, it does not extend to contexts of unethical collaboration. They also refute common concerns that partisan homogeneity may facilitate cooperative corruption.
Can monuments to victims of authoritarian regimes promote more tolerant societies? We look into the case of Stolpersteine, small memorials commemorating victims of the Nazi regime in Germany. Unlike other monuments, Stolpersteine are dedicated to specific individuals who were victims of Nazi violence. In a pre-registered survey experiment, we showed treated individuals pictures of the stones. Our results show that exposure to Stolpersteine strongly increased negative emotions and reduced positive ones. The results on attitudinal and quasi-behavioral outcomes are mixed, likely driven by ceiling effects. We find a positive effect on tolerance toward the only minority group that faces low tolerance in the control group. However, there is no statistically significant effect for other groups.
The partisan gap in economic perceptions flipped unusually dramatically after the 2024 U.S. presidential election: following the Republican victory, Democrats (Republicans) suddenly rated the economy much more negatively (positively). Was the resulting partisan difference a case of expressive responding, wherein surveys exaggerate partisan bias in measures of economic perceptions? In April 2025, I fielded a panel survey experiment that asked survey respondents to guess then-unpublished measures of economic growth, inflation, and unemployment in the current month or quarter (Prolific, N = 2,831). Randomly selected respondents were offered $2 per correct answer. Partisan bias did not shrink as a result, suggesting genuine differences in economic perceptions. Two measures of response effort (response time and looking up answers) increase, suggesting that misreporting does not fully explain the effects of pay-for-correct treatments.
Scholarship has identified key determinants of people's belief in misinformation predominantly from English-language contexts. However, multilingual citizens often consume news media in multiple languages. We study how the language of consumption affects belief in misinformation and true news articles in multilingual environments. We suggest that language may pass on specific cues affecting how bilinguals evaluate information. In a ten-week survey experiment with bilingual adults in Ukraine, we measured if subjects evaluating information in their less-preferred language were less likely to believe it. We find those who prefer Ukrainian are less likely to believe both false and true stories written in Russian by approximately 0.2 standard deviation units. Conversely, those who prefer Russian show increased belief in false stories in Ukrainian, though this effect is less robust. A secondary digital media literacy intervention does not increase discernment as it reduces belief in both true and false stories equally.
Scholars increasingly conceptualize populism by whether politicians use people-centric and anti-elite appeals that pit a homogeneous people against a corrupt elite. These appeals reflect "thin" ideology because they offer no programmatic content and thus politicians must pair these appeals with more substantive positions, termed their "host" (or thick) ideology, which often consists of nativism on the right (e.g., espousing anti-immigrant positions) and socialism on the left (e.g., prioritizing redistribution). An emerging literature has thus sought to estimate whether populists garner support due to their thin ideology or their substantive host ideology. To date, no research has validated whether populism treatments (1) truly operationalize populist thin ideology, and (2) do so without manipulating host ideology. Results from three conjoint validation experiments fielded in both the United States and the United Kingdom show that thin ideology treatments successfully manipulate the underlying concepts but caution that some operationalizations also affect perceptions of host ideology.
Solar geoengineering offers a speculative means to cool the planet by reflecting solar radiation into space. While some research suggests that awareness of solar geoengineering could reduce public support for decarbonization through a moral hazard mechanism, other studies indicate that it could serve as a "clarion call" that motivates further action. Using a pre-registered factorial design, we assess how sharing balanced information on solar geoengineering affects attitudes toward decarbonization policies and climate attitudes among 2,509 US residents. We do not find that solar geoengineering information affects support for decarbonization on average, though it may increase support among initially less supportive subgroups; moreover, this information tends to increase the perception that climate change is a daunting problem that cannot be resolved without decarbonization. Our results suggest that concerns about moral hazard should not discourage research on solar geoengineering - as long as the public encounters realistic messages about solar geoengineering's role.
As digital platforms become a key channel for political advertising, there are continued calls for expanding regulation of digital political ads as a distinct content category. However, designing policies to meet these demands requires us first to decipher what the public perceives a "political" ad to be. In this article, we report two preregistered experiments to understand factors that drive public perceptions of what makes an ad political. We find that both advertiser-level cues and content-level cues play an independent role in shaping perceptions. To a lesser extent, participants also attribute political meaning to ads that clash with their own preferences. These patterns were replicated in a conjoint study using artificial ads and in an experiment using real-world ads drawn from the Facebook Ad Library. Our findings serve as an important benchmark for evaluating proposed definitions of political ads from policymakers and platforms.
This study examines whether Americans are more supportive of immigration when migrants share their partisan preferences. To address this question, we embedded a preregistered experiment in a nationally representative survey that was fielded the week before the 2024 US Presidential Election. The main experimental treatment provided information that some immigrant groups tend to favor Donald Trump and the Republican Party. This information reduced support for immigration among Democrats and increased support for immigration among Republicans. Our findings suggest that immigrants' political identities impact public support for immigration. They also suggest that Trump's apparent gains among immigrant voters in the 2024 election have the potential to reduce partisan polarization over immigration in the future.
Despite general public support, efforts to build affordable housing often encounter stiff resistance due to "not in my backyard" (NIMBY) attitudes, which are often rooted in false or unsupported beliefs about affordable housing and its impacts on surrounding communities. Would correcting these misperceptions increase support for building affordable housing? To answer this question, we conducted a preregistered survey experiment measuring how support for affordable housing in the U.S. varies at different distances from where respondents live (one-eighth of a mile away, two miles away, or in their state). Our results indicate that correcting stereotypes about affordable housing and misperceptions about its effects increase support for affordable housing. Contrary to expectations, these effects are often larger for affordable housing near the respondent's home (rather than at the state level), suggesting that debunking myths about affordable housing may help to counter NIMBY attitudes.
