Experimental Lab - Publications

Author: Madeline C. Hornyak

The effectiveness of persuasive messaging in American political campaigns has been a topic of research for many years (Lenz, 2024), particularly due to the fact that political campaigns must take product marketing concepts, political communication techniques and combine them to market a candidate and the party as a whole as a product (Lees-Marshment, 2001). Political marketing is difficult due to the “several, possibly conflicting markets,” meaning that campaigns must market to several different demographics and win over a majority of voters that may have nothing in common and have differing factors that affect their voting behaviors (Lees-Marshment, 2001).

This study looks to analyze the political messaging by male and female candidates in candid platforms like Twitter and how that messaging affects voter perceptions and voter behavior, as well as how the gender of political candidates changes the way voters respond to social media posts. Studies have typically focused on formal campaign messaging from speeches to interviews and television advertisements, but there is little up-to-date research on the effects of social media on voter perceptions and voting behaviors, as social media is constantly growing and evolving, thus changing the landscape of political marketing. According to Terri L. Towner and David A. Dulio (2013), with the rise of digital platforms since the 2008 presidential election, social media has become a critical space for candidates to engage directly with voters, giving them the ability to bypass traditional media outlets. This study will be conducted through an analysis of six senate races across the country in the 2024 election, based on the examination of actual candidate Tweets disseminated during the 2024 senate campaigns, as well as a study performed on college students. By focusing on a younger demographic, the study aims to provide insights into how social media influences not only voting behaviors but also broader political engagement in the digital age.

College students represent a critical voting demographic, as they are at a stage in life where they are still forming political opinions that can be shaped by new information and experiences. It is well understood that college is a time that shapes young people’s political attitudes (Frenz, 2020). Additionally, this demographic is the most active on social media platforms such as Instagram, X (formerly Twitter), TikTok, and Facebook (Pew Research Center, 2024). Furthermore, young voters have historically had lower turnout rates compared to older demographics, so this study aims to understand the factors that motivate or discourage young voters from participating in elections (Hartig et al., 2023). Understanding the impact that social media plays in political discourse can offer valuable insights for political campaigns and civic engagement groups. Understanding how college students respond to political messaging on platforms like X (Twitter) can help candidates tailor their outreach strategies to foster greater political participation among these voters.
 

Author: Lydia Dawson

Presented at the 2026 WPSA Conference

Abstract: Do Instagram comments impact voters’ perceptions, and are there gendered differences? Building on literature exploring the role of gender stereotypes in political contexts, I hypothesize that respondents will evaluate female incumbents more harshly than male incumbents after viewing negative Instagram comments associated with them. I conducted an experiment where 436 undergraduate students were randomly assigned to one of four experimental groups. Students were assigned a hypothetical male or female incumbent and shown identical candidate biographies. Two groups were presented with negative comments associated with the candidates (consistent between groups). Each group was asked to evaluate their candidate’s viability, favorability, competency, leadership ability, qualifications, and the likelihood that they would vote for them. Experimental groups completed these evaluations as both a pre- and post-test. I found that the male candidate scored lower than the female candidate on all dependent variables, both before and after stimulus presentation. Both candidates were subject to lower evaluation scores after presentation of the comments, with scores for the female candidate dropping by a slightly larger margin, though remaining higher than those of the male candidate. Voters’ perceptions, then, are impacted by Instagram comments regardless of a candidate’s gender, though there is some evidence to suggest perceptions of female candidates may be more negatively affected.

Author: Keira Hammond

This paper was presented at the 2026 Annual General Meeting for the Southwest Social Science Association in Nashville, TN.

Abstract: This study examines whether personality traits are directly associated with support for limited government or whether their effects operate indirectly through ideological orientation. Focusing on conscientiousness, the analysis tests a mediation model across two datasets: an original survey of 372 undergraduate respondents and a nationally representative sample from the 2016 American National Election Study (ANES). Across both studies, conscientiousness is modestly associated with policy preferences at the bivariate level, but this relationship is substantially reduced once ideological variables are included, with ideology emerging as the strongest predictor of mass policy preferences. Mediation analyses reveal a consistent indirect effect of conscientiousness through conservatism across both samples. In the ANES data, this relationship varies across domains: the effect is partially mediated for social policy attitudes, with a small direct effect remaining, and fully mediated for economic policy attitudes. The indirect effect is significantly stronger for social than for economic policy attitudes. These findings suggest that personality is associated with political behavior primarily through ideological orientation, while also exhibiting domain-specific variation.
 

Author: Namig Abbasov (Stanford University) and Xiao Hu (University of Arizona)

Presented at The 26th IEEE International Conference on Advanced Learning Technologies

Abstract: While institutional guidelines for AI use are expanding, there is limited experimental evidence on how students evaluate different forms of AI integration in course design. This study examines how alternative AI policy configurations affect perceived usefulness, support for adoption, and autonomy. Using a conjoint experiment in which students evaluate hypothetical course scenarios varying AI use across instructional materials, grading, and personalized support, we causally estimate the effects of specific design choices. Results show that evaluations are driven primarily by human oversight rather than AI use per se: systems operating without instructor review—especially in grading—produce large declines in all outcomes, whereas AI
with oversight is evaluated similarly to fully human instruction. These findings demonstrate that student acceptance of educational AI depends on governance design, identifying human oversight as a key determinant of perceived legitimacy.
 

Author: Matthew C. Dempsey and Joshua R. Thompson

Journal: Social Science Quarterly 2026

Abstract: Objectives. To test whether cues to a judge’s sexual orientation influence perceptions of judicial impartiality and whether such effects have changed over time. Methods. Two between-subjects survey experiments conducted nine years apart (N = 836 in 2014; N = 493 in 2023) asked respondents to evaluate a state-court judge described with heterosexual, gay or lesbian, or no-orientation cues. Identical vignette structures enabled direct comparison across contexts. Results. In 2014, respondents rated gay and lesbian judges as more biased than heterosexual judges, while perceptions of correctness showed little difference. By 2023, orientation-based bias had mostly receded: ratings of gay male judges no longer differed from heterosexual or no-information judges, and only a modest residual bias toward lesbian judges remained. Conclusions. Even in a more accepting social climate, subtle identity cues can still shape perceptions of impartiality, refining conditional-legitimacy theories of judicial evaluation.

Author: Matthew C. Dempsey

Award: Best Paper Award, Southwestern Political Science Association, Southwestern Social Science Association Annual Meeting, Nashville, TN. (April 2026)

Abstract: As political campaigns increasingly deploy artificial intelligence (AI) to generate communications, 26 U.S. states have enacted AI disclosure laws. This study tests how AI authorship disclosure affects candidate evaluations through a 2×2 experiment (message medium × disclosure type, N=518) with a biographical control. Drawing on dual-process theory and algorithm aversion research, I predicted AI disclosure would reduce authenticity and trust through pathways consistent with heuristic (System 1) rather than systematic (System 2) processing. AI-disclosed messages reduced candidate evaluations relative to human-disclosed messages (d=0.24–0.52 across conditions and outcomes) consistently across Twitter posts and fundraising emails. Mediation analyses revealed approximately 68% of the total effect operated through heuristic pathways, reduced authenticity perceptions (~39%) and reduced trust (~29%), while message quality perceptions showed no significant mediation, a pattern consistent with heuristic rather than systematic processing, though the design does not directly manipulate processing depth. Effects were robust to demographic controls, replicated across two universities, but moderated by campaign AI attitudes. Findings validate psychological mechanisms underlying state disclosure requirements and demonstrate authenticity concerns, not quality judgments, drive voter responses to AI-generated campaigns.

Author: Justin Michael Zyla

Journal: American Politics Research (2024)

Abstract: Imagine two incidents of political violence. In the first, you share political affiliation with the victim. In the second, they reside in the opposite party. How would this minor change – a shifting label, the difference of a word – impact your reaction? This article offers empirical insight through an experiment: U.S. participants read a mock college controversy, where a student sent death threats to, and doxed, a professor. The treatment varied whether the perpetrator described the professor as a Democrat, Republican, or used otherwise non-descript (e.g., “political”) adjectives. A posttreatment survey then measured respondents’ discrete emotions, the penalties they preferred the student receive, and their partisan group identity strength. Participants who read about violence against a copartisan victim showed a statistically significant increase in preferred penalty severity. But violence against an outparty victim mirrored the control, with subjects reacting as if they didn’t know the political affiliation of anyone involved. Posttreatment measures also demonstrated the potential for anxiety (but not anger or partisan strength) to mediate this underlying partisan bias.

Authors: Kim Fridkin, Trudy Horsting, Anastasia Brown, Alexandra Williams

Journal: American Politics Research (2023)

Abstract: We explore how political and psychological factors condition the effectiveness of PSAs promoting COVID-19 vaccines. Targeting college students, we utilize a pretest-posttest experiment to examine how different PSAs (emotional, informational, and humorous) influence students’ emotional reactions and assessments of the PSAs. Further, we assess whether the PSAs are able to influence learning and persuasion. We find certain PSAs are more effective at changing people’s attitudes about the COVID-19 vaccine and the impact of these messages depends on people’s political and psychological predispositions. The informational PSA produces learning, regardless of students’ receptivity to pro-vaccine messaging. However, the humorous and emotional PSAs encourage learning only for those who are already receptive to the vaccine. These findings have implications for future public health campaigns aimed at college students, suggesting PSA campaigns developed to battle new health crises should be launched quickly before people develop strong attitudes about the emerging crisis.

Author: Kenicia Wright

Presentation: 2023 Women in Legislative Studies Conference, hosted by Brown University

Abstract: This paper aims to make a two-fold contribution to scholarly understanding by focusing on how young potential voters assess candidates and the factors that explain their willingness to donate to hypothetical candidates. On the one hand, in recent elections, younger voters (18-25 years old) have turned out in impressive numbers, particularly in battleground states, and may continue to play influential roles in shaping future electoral outcomes, yet studies that focus on the voting behavior of young voters in the US are rare as most existing research studies this topic by comparing young voters to older voters, young voters in the US to those in other countries, or exploring the effects of electoral rules. On the other hand, early theories on the electoral success of women often focus on the effects of a single identity, Congressional Elections, and posit minority women as being “doubly disadvantaged” due to their race/ethnicity and gender. Increasingly diverse pools of candidates running – and winning – elections, particularly for state legislatures. Campaign contributions are often emphasized as critical in shaping electoral success, yet an expanding body of intersectionality research suggests that minority women have difficulty generating campaign donations compared to other candidates. What explains the decision to donate to a candidate by young voters? Are these factors different than the factors important in shaping their vote choice? I argue that the proclivity for representation will interact with electoral viability to explain this decision and that the factors important in shaping these perceptions will vary based on multiple, overlapping characteristics of candidates. Using a randomized survey experiment of young potential voters from a pilot study fielded in 2022, preliminary results provide mixed support for expectations - voters are more likely to donate to candidates as the shared characteristics between them increases, though not all characteristics have the same impact. This project highlights the need for more theories that focus on this potentially politically influential, understudied group of voters and reassessments of prevalent theories in light of changing dynamics in political contexts in the US.

Authors: Peyman Asadzade and Roya Izadi

Journal: Research & Politics (2022)

Abstract: Large-scale military aggression is argued to damage the international image of the aggressor and mobilize global public opinion against it. Previous cross-country research also finds that negative views of the aggressor are usually limited to the government and do not extend to the citizens of the invading country. Our article provides micro-level evidence on attitude change toward Russia as a country, the Russian people, and the Russian government after its invasion of Ukraine. We use data from a survey conducted between the morning of 21 February 2022 (3 days before the Russian invasion of Ukraine) and the night of 28 February 2022 (5 days after the invasion) in the United States to evaluate how the Russian invasion of Ukraine affected attitudes toward the country, its people, and the government. We also conduct a subgroup analysis to explore the magnitude of attitude change across sociodemographic and political subgroups after the invasion. Our findings show fairly significant damage to the image of Russia as a country as well as the Russian government. However, the reputational damage of the Russian people is minimal. The results also suggest that Republican and religious subgroups had the largest attitude change on Russia and the Russian government.

Author: Angguntari Ceria Sari

Presentations: Dissertation presented at  2020 ISA Mid-West Conference, 2021 ISA West and Annual Conference, and 2022 Midwest Political Science Association Conference

Abstract: How do political elites perceive regional intergovernmental organizations that seek to promote democracy? When do political elites view regional intergovernmental institutions promoting democracy as legitimate? Membership in a regional organization that promotes democratic values has a positive impact on democratization or democratic consolidation. It can improve domestic political support for the democratic regime. The leaders of the new regime will receive domestic backlash if they engage in anti-democratic activities and ruin the chance at global acceptance. This approval from international or regional organizations can also encourage the population to continue the transition or democratic consolidation process. The underlying assumption of the abovementioned argument is that people hold favorable views of regional democracy organizations such that they honor the commitments to political reform. Whether political elites have confidence that the regional organization can positively impact democratic consolidation and how they perceive different types of regional organizations promoting democracy are empirical questions. This project seeks to answer these questions through a small sample and non-population-based elite survey experiment in Malaysia, Indonesia, and the Philippines and a laboratory survey experiment with a larger sample of university students, which gives me the opportunity for statistical power. The results suggest that political elites are skeptical of regional intergovernmental bodies promoting democracy. Meanwhile, non-elites consider regional institutions promoting democracy illegitimate when they are informal, i.e., no written shared expectations, rules, and permanent secretariate. When regional interstate cooperation on democracy operates under a formal procedure or codified in an international treaty and supported by a permanent secretariat, non-elites tend to consider them more legitimate.

Author: Namig Abbasov

Presentations: Dissertation presented at the International Studies Association Midwest

Abstract: Why do states repress sexual minorities? The law of coercive responsiveness posits that states resort to repressive measures against groups they view as a threat. However, despite their limited potential to undermine state security and the negative global ramifications, sexual minorities are frequently subjected to repression. This study proposes a link between state repression of sexual minorities and opposition to the US-led liberal international order: States opposed to this order use the repression of sexual minorities as a political tool to achieve domestic objectives, such as cracking down on domestic pro-Western forces and diverting public attention from poor economic performance. State repression of sexual minorities draws international attention, with Western leaders issuing condemnations and calling for sanctions. These condemnations and calls for sanctions allow anti-Western authoritarian regimes to link homosexuality to the West and frame it as a "Western-imported" and "alien". These narratives appeal to the anti-Western authoritarian countries' domestic audience and establish negative perceptions of domestic pro-Western opposition, making it less costly for political leaders to discredit and target domestic pro-Western forces. Political regimes also frame Western reactions as "intervention to national sovereignty" and "imposition of moral distortion upon our country," raising nationalist sentiments, which help them to obscure demands for redistributive policies and neutralize possible social unrest and civil disobedience. This study utilizes new data on state repression of sexual minorities and reports significant findings consistent with the argument. To assess the model performance, the study employs machine learning techniques such as random forests and neural artificial networks and finds that the models based on the opposition to the US-led liberal order explanation display high accuracy in predicting state repression of sexual minorities.

Authors: Fabian G. Neuner and Michael Hechter

Presentation: 2021 ISPP

Abstract: There is a large literature in the social sciences and psychology suggesting that individuals who suffer from a reversal of social status experience negative affect and, some aver, intolerant attitudes towards the groups that threaten to replace them in the status hierarchy. Political scientists and sociologists have long attributed right-wing politics to the loss of status among previously advantaged groups, and this view has persisted to the present day. This paper adds an important caveat to this line of research. The paper seeks to identify the causal impact of actual status loss on in-group bias, outgroup discrimination, and potential downstream political attitudes experimentally. Crucially, the experiments isolate the effect of status reversal from that of pre-existing beliefs about lower status outgroups that may confound the relationship. It suggests that status reversal may only have the posited effects given a history of discrimination against specific outgroups. We therefore conclude that status reversal is a necessary but insufficient condition for the development of intolerant attitudes and behaviors. The sufficient conditions for the development of reactionary politics likely consist of both actual status reversal and pre-existing intergroup stratification.

Authors: Madison Andrade

Publication: Barrett, The Honors College Thesis/Creative Project Collection 

Abstract: This study aims to gain a better understanding of the potential influence that perceived levels of attractiveness impose on individuals’ perceptions of male and female candidates’ competency and favorability. I utilized a quasi-experimental design that had a total of 603 respondents, all of which were current students at Arizona State University. The experimental design of the study consisted of the same pre-test, and post-test questionnaire, with a randomly assigned experimental group viewing a video regarding the halo effect, and a randomly assigned control group viewing a neutral video with no comment on attractiveness and its inherent effects. The results demonstrate insignificant findings in regard to candidate favorability and attractiveness, but one significant finding revealed itself in regard to competency. Hypothesis 6 (H6) identified an unexpected, yet statistically significant finding when concerning the ‘above average’ female candidate and respondents’ perception of her competency. This suggests that individuals may be more inclined to favorably perceive the competency of an attractive female candidate when they are aware of their implicit biases that accompany increased levels of attractiveness. This study offers insight into whether perceived level of attractiveness affects evaluations of male and female politicians. 

Authors: Babak RezaeeDaryakenari and Peyman Asadzade

Journal: Research and Politics (2020)

Abstract: The color revolutions in Eastern Europe and the recent waves of protests in the Middle East and North Africa show that nonviolent conflicts, like violent conflicts, spread across countries. Scholars argue that the idea of adopting nonviolent resistance travels across borders because dissidents learn from foreign protestors about the principles of nonviolent campaigns and their success in changing the status quo. However, the commonly used country-level data do not allow scholars to distinguish between these two learning mechanisms and to establish a causal association between exposure to information about nonviolent resistance and changing individuals’ attitudes. Therefore, we conduct an experimental survey to analyze and distinguish individuals’ responses to information about the principles of nonviolent campaigns and their higher prospects for success. Our findings show both types of information increase individuals’ evaluation of nonviolent resistance. However, exposure to information on the principles of nonviolent campaigns yields different results than exposure to information on the success rate of nonviolent campaigns.

Author: Mark Ramirez

Journal: Results from this study are included in Mark Ramirez and David A.M. Peterson's book Ignored Racism: White Animus Toward Latinos. New York. Cambridge University Press (2020)

Abstract: Although Latinos are now the largest non-majority group in the United States, existing research on white attitudes toward Latinos has focused almost exclusively on attitudes toward immigration. This book changes that. It argues that such accounts fundamentally underestimate the political power of whites' animus toward Latinos and thus miss how conflict extends well beyond immigration to issues such as voting rights, criminal punishment, policing, and which candidates to support. Providing historical and cultural context and drawing on rich survey and experimental evidence, the authors show that Latino racism-ethnicism is a coherent belief system about Latinos that is conceptually and empirically distinct from other forms of out-group hostility, and from partisanship and ideology. Moreover, animus toward Latinos has become a powerful force in contemporary American politics, shaping white public opinion in elections and across a number of important issue areas - and resulting in policies that harm Latinos disproportionately.

Author: Fabian G. Neuner

Presentation: iPoster Presentation at 2020 APSA

Authors: Kim Fridkin, Patrick Kenney, Bartia Cooper, Ryan Deutsch, Manuel Gutierrez, and Alexandra Williams

Journal: Political Research Quarterly  (2020)

Abstract: We compare two alternative measures for assessing people’s emotional reactions to political stimuli: the traditional self-report measure and facial expression analysis. We recruited participants to take part in a study examining reactions to a set of negative political commercials aired during the 2018 elections. We compare people’s self-reporting of their emotional reactions to negative political advertisements with their expressed emotion, according to the facial expression analysis. We find the discriminant validity of the facial expression analysis is higher than the self-report measure. Second, the self-report and facial expression measures of emotion have little convergent validity: we fail to find a consistent and strong positive correlation between the self-report and facial software measures of the same emotion and the same political advertisement. Third, the facial software measure has better predictive validity than the self-report measure, generating better predictions for the three dependent variables examined: changes in political interest, changes in people’s confidence in elected officials, and people’s assessment of the tone of the senate campaign.

Authors: Michael Chacon and Amanda McCubbins 

Presentation: Poster Presentation at the Experimental Social Science, NYU 2019. This project began in POS 503

Authors: Gina Serignese Woodall, Richard Herrera, Joshua R. Thompson, and Jorge Coss Ortega 

Journal: Journal of Political Science Education (2017)

Abstract: Summer bridge programs are supposed to connect a graduating high school senior’s summer to their first semester in college, easing the transition away from home and into a university setting. Although research is plentiful on the programs, assessments regarding the overall effectiveness of such programs have been mixed (e.g., Cabrera, Miner, and Milem 2013; Douglas and Attewell 2014; Grayson 2003). As Cabrera, Miner, and Milem (2013) note, many of the studies collect data from participants in a one-time satisfaction survey and/or do not have an equivalent group of students who did not participate in the program from which to compare effectiveness. Our proposed study mitigates these flaws by employing a nonequivalent-groups quasi-experiment (NEG). Our bridge program—called Early Start (ES)— tapped into social, emotional, and academic engagement, as we sought to integrate the students into the university and school communities while holding high expectations and actively involving them in their own learning (Tinto 2004). Results suggest that not only were students enrolled in our program better socially and emotionally integrated into the university during their first year but they also scored better on content-based political science questions at the end of the fall semester, compared to their non-Early Start peers. Finally, aggregated data indicate an increase in retention for Early Start participants.

Authors: Gina Serignese  Woodall and Tara M. Lennon

Journal: American Political Science Association (2017)

Abstract: How can educators harness social media to promote civic engagement? Over three spring semesters, we conducted a nonequivalent control group design experiment in which one class each semester was required to tweet weekly about the course topics and another class was not. In the third year, we also conducted focus groups about students’ use of social media to obtain and share political information. Preliminary results of the experiment show that the treatment classes had statistically higher political knowledge and political engagement through Twitter compared to the control classes. Preliminary results also suggest some nuanced gender differences in how participants use Twitter. Compared to other social media, such as Facebook, Twitter leans toward more direct, personal—and often confrontational—communication. While female participants used Twitter for political purposes to a greater extent than male students in our study, the gendered impact of the Twitter requirement is not clear. Our focus group discussions further developed some of our findings regarding the factors that motivate students to follow certain types of political social media accounts. By examining students’ motivations and social media behaviors, this chapter contributes to the emerging research on how educators can encourage e-civic engagement habits that foster active citizenship and political knowledge.

Author: Amanda Wintersieck

Journal: American Politics Research (2017)

Abstract: How does fact-checking influence individuals’ attitudes and evaluations of political candidates? Fact-checking plays an increasingly important role in U.S. elections, yet we know little about its impact on voters. To answer this question, I conduct an experiment utilizing one political debate from the 2013 New Jersey Gubernatorial race in combination with fact-checks that offer either confirming information or corrective information to determine how fact-checking influences evaluations of candidates. I find that evaluations of the candidate’s debate performance and evaluations of the debate winner are improved by the presence of a fact-check that confirms the accuracy of a candidates’ statement and lowered by fact-checks that state that a candidate is being dishonest. Moreover, respondents indicate greater willingness to vote for a candidate when the fact-check indicates that the candidate is being honest. These findings suggest that fact-checking can influence people’s evaluations of political events.