PEER-REVIEWED PAPERS
MEASURING BELIEF CERTAINTY IN POLITICAL KNOWLEDGE. Political Behavior. 2024.
ON THE INDEPENDENT ROLES OF COGNITIVE & POLITICAL SOPHISTICATION: VARIATION ACROSS ATTITUDINAL OBJECTS
ON THE INDEPENDENT ROLES OF COGNITIVE & POLITICAL SOPHISTICATION: VARIATION ACROSS ATTITUDINAL OBJECTS (with Joe Vitriol, Joe Sandor, and Christina Farhart). Applied Cognitive Psychology. 2022.
Political knowledge measures what people know and how they reason about politics, but scholars still debate how to best measure the concept. A growing body of survey research has demonstrated the theoretical importance of measuring an individual’s expressed belief certainty about their own informedness. Yet existing political knowledge measures typically fail to recover this information. This article validates a question format that allows respondents to express their level of certainty in their knowledge about politics and government. The analyses show that the certainty-in-knowledge format: (1) reduces systematic differences between men and women in responding “Don’t Know,” and (2) decreases answer look-up on online surveys. Validity, reliability, and information analyses further demonstrate that incorporating belief certainty results in a knowledge scale that displays theoretically expected relationships with a range of outcome variables while also having superior psychometric properties. Additionally, the certainty-in-knowledge format eliminates the confounding influence of response option cues and can be applied to questions on any topic—political figures, institutional rules, policies, and surveillance facts. This shift in measurement can be made without costly changes to survey questionnaire design and length. Finally, the new format represents a more realistic conception of citizen competence: one that elevates the ability of discerning true from false political information.
People are motivated to maintain consistency between importantly held identities, preferences, and judgments. In political contexts, motivated reasoning can help explain a wide range of political phenomena, including extremism, polarization, and misperceptions. However, recent findings in psychology have challenged this account. These perspectives emphasize the role of cognitive sophistication (e.g., analytical reasoning, numerical literacy) in political attitudes, but differ in terms of whether it is expected to attenuate or exacerbate politically motivated reasoning and belief in conspiracy theories. Yet prior investigations have not examined the relative and independent effects of both political and cognitive sophistication. Using data from two samples, including one sampled to approximate representativeness in the U.S., we demonstrate that both types of sophistication have independent and (at times) countervailing effects on belief in COVID-19 conspiracy theories and other political attitudes. Our results are critical for theories of cognitive sophistication, political cognition, and attitudes, and the psychology of conspiracy theories.
ISSUE IMPORTANCE AND THE CORRECTION OF MISINFORMATION (with Jenn Jerit). Political Communication. 2022.
The study of misinformation—and its correction—has proliferated in recent years. Yet the empirical record includes instances where corrective messages do and do not work, even on similar issues. Despite intense scholarly attention to this topic, it remains unclear when people will revise false beliefs. Our study examines a factor with a long history in the study of public opinion: the importance a person attaches to an issue. The subjective state of issue importance has complex effects. It can increase an individual’s motivation to engage in effortful information processing while also leading them to defend existing beliefs and opinions. In a series of experiments administered in national surveys, we examine whether issue importance is implicated in the failure to correct false beliefs. The analyses show that on the topic of GM foods, the effects of a corrective message are smallest among misinformed people who rated the issue as personally important. By contrast, framing GM foods in terms of partisan identity engendered little resistance to a corrective message. Our findings illustrate the value of adopting a broader perspective on misinformation because people may resist corrections for reasons that are unrelated to their partisanship.
AUTHORITARIANISM AND RIGHT-WING VOTING IN BRAZIL. Latin America Research and Review. 2022.
Mounting evidence shows that authoritarian orientations exert a powerful influence on public opinion attitudes and candidate support. The 2018 Brazilian elections brought to power Jair Bolsonaro, a candidate with an open disregard for democracy and democratic institutions. This study examines Brazilian voters’ differences in authoritarianism and electoral support for a right-wing authoritarian candidate. Employing the AmericasBarometer national survey data, results demonstrate that authoritarianism is politically important in Brazil due to its association with attitudes toward the use of force as well as with social and political conservative attitudes. The effect of authoritarianism on the probability of voting for Bolsonaro is as large as that of other relevant political behavior variables, such as ideology, negative partisanship, or religiosity, whereas nonauthoritarian voters spread their votes across other candidates. Although these other variables are also relevant to Bolsonaro’s victory, his candidacy was uniquely able to mobilize a coalition of authoritarian voters. Whether or not authoritarianism remains a salient cleavage in the electorate is considered along with the consequences of this potential divide for political competition in Brazilian politics.
AFFIRMATIVE ACTION ATTITUDES OF WHITES: EVIDENCES FROM A LIST-EXPERIMENT SURVEY IN BRAZIL. Brazilian Political Science Review. 2018.
Recently in Brazil, public policies began to be implemented to reduce discrimination and promote individuals that have been excluded due to a personal characteristic: race. However, public opinion on such racial policies is not consensual, especially among white individuals. In this paper, I examine the opposition to affirmative action among college students and ask: are these attitudes stemming from negative racial affect, conflict between social groups or individual political predispositions? A list-experiment is used for its potential to offset the social desirability bias, as it allows respondents to be indirectly questioned, ensuring greater privacy in their answers, and allowing more sincere opinions. A survey collected students’ opinions at a large Brazilian university, and results shows that social desirability bias is strong among white students in Brazil. Indeed, while 40% of white respondents showed approval of the policy when asked directly, only 6% of them approve it sincerely. Thus, close to 34% respondents over-reported their true preference. Moreover, I also estimate the individual determinants of affirmative action support and I find that more knowledgeable students show greater support for the racial quotas. Finally, results indicate that both negative racial affect and political predispositions explain white students’ opposition to affirmative action policies.
BOOK CHAPTERS
PROGRESSIVE POLICY CHANGE, CULTURAL BACKLASH, AND PARTY POLARIZATION IN LATIN AMERICA. In: The Recasting of the Latin American Right. Cambridge University Press. (with André Borges).
WHEN, HOW, AND WHY PERSUASION FAILS: A MOTIVATED REASONING ACCOUNT. In: The Oxford Handbook of Electoral Persuasion. 2020. (with Ryan Cotter and Milton Lodge)
What drives recent party system polarization in Latin American democracies? This research seeks to understand the recent rightward move in many party systems in the region and the ensuing party system polarization processes. Some of the countries that avoided polarization in the twentieth century are today confronted with much more polarized party politics than those that experienced it early on. We discuss different explanations that focus either on long-term economic distress or on cultural backlash, and we propose an alternative explanation that focus on the strategic behavior of right-wing parties and candidates in response to progressive policy changes on issues such as LGBT rights. We argue that where a left-wing president was in power and progressive policy changes were enacted, right-wing challengers had an easier time in appealing to voters dissatisfied with the incumbent’s policies and in building a Manichean, polarizing narrative. In contrast, where the incumbent was a right-leaning or centrist party that obtained success in pushing forward a progressive policy agenda (or progressive changes were implemented due to supreme court decisions), it was much harder for an opposition right-wing party to construct a polarizing narrative. This proposition is tested using data on Latin American parties’ ideological positions and electoral returns covering 16 countries. Our statistical results do not provide support for either economic explanations or demand-driven, cultural backlash accounts. Instead, we find that polarization increases when leftist incumbents are associated with progressive policy change. We also find that right-wing parties become more ideologically extreme in this latter scenario.
All citizens, including the most sophisticated and attentive, possess a powerful drive to defend their opinions, attitudes, and worldviews in the face of challenges. This bias manifests as an active effort to rationalize what they want to believe rather than accepting incongruent viewpoints. This propensity to dismiss or counterargue perspectives we do not like is evident across political identities and raises serious obstacles to successfully shifting others’ attitudes. This article presents an integrated understanding of these information processing phenomena under the John Q. Public model and explains how prior beliefs, confirmation bias, and disconfirmation bias interact to produce persistence in evaluations. It then goes on to explain how network analysis and experimental designs can be leveraged to illuminate the black box of communication effects and deepen our understanding of when persuasion is successful. Mapping out the cognitive relationships between semantic concepts and connecting these to a raw, affective charge provides a clearer view of citizens’ understanding of issues and how their thoughts and feelings shift in response to targeted political messaging.
THE BOUNDARY CONDITIONS OF MOTIVATED REASONING. In: The Oxford Handbook of Electoral Persuasion (with Ryan Cotter and Milton Lodge). 2020.
Decades of research on motivated reasoning have found that citizens routinely place a higher priority on defending their preexisting beliefs than on updating them in response to new and conflicting experiences. Scholars have employed a variety of strategies in their efforts to eliminate, or at least mitigate, the reactive and polarizing effects of motivated reasoning. In this chapter, we provide a conceptual framework for navigating this body of literature that categorizes these efforts based upon how their interventions moderate the three central dimensions of information processing: priors, motivations, and considerations. This review of an extensive and growing literature finds that purely cognitive approaches to political communication are insufficient for overcoming subjects’ powerful drive to rationalize their own perspective. Accepting, understanding, and embracing the role of affect in shaping information processing offers promise for redirecting attention and reshaping intentions for the purpose of overcoming the limitations of situated perspectives.
SELECTED PAPERS IN PORTUGUESE
Lesser-of-two-evils? Partisan Identity and Ambivalence in the Brazilian Electorate. Opinião Pública (with Álvaro Pereira).
“From Lulismo to Anti-petismo? Polarization, partisanship and voting in Brazilian presidential elections.” Opinião Pública (with André Borges). 2018.
Research on partisan voters in Brazil have demonstrated a decline in both partisanship and feelings toward political parties. How is party identity maintained in this situation? We affirm that even when voters do not evaluate their own party positively, their party identification persists through their interparty feelings, which have deteriorated in recent years. Data from the Brazilian Election Studies (ESEB) for the period from 2002 to 2018 suggest that partisan voters are less enthusiastic about Brazilian political parties, and the relationship between partisan feelings has strengthened in the most recent electoral context. The results about the relationship between party sentiments, to use a quadratic term, indicate that the main alternatives to support for the Worker’s Party (PT) justified a
party identity based on the ‘lesser of two evils’ strategy. This indicates the use of a greater deterioration of the image of a political opponent as a mechanism to justify ambivalence in relation to
their own partisan position.
The recent debate on party identification and voting in Brazil has pointed to the growing importance of voters’ positioning in relation to the two main presidential parties – PT and PSDB – on the choice of presidential candidates. In this paper, we contribute to the literaturequestioning some diagnoses related to the consolidation of the Brazilian presidential party system. In particular, we point to the methodological and theoretical fragility of analyzes that see PT-PSDB polarization as a reflection of a growing split of the electorate into two clearly differentiated and polarized groups. The results of the descriptive analysis and multivariate statistical models using the the Brazilian Electoral Study (ESEB) conducted in 2002, 2006, 2010 and 2014 show that, despite the growing importance of party sentiments in determining the voters’ electoral behavior, there is no evidence that such movement is associated with increased ideological mass polarization. Actually, we observe that the ideological and attitudes differences between PT and PSDB supporters are small, and there is actually ideological convergence between the various electorate segments.
EDITED VOLUMES
Despite the growing interest in the study of social movements, political parties, activists, and right-wing voters, there has been a noticeable absence of work capable of integrating various approaches and offering a conceptual and empirical synthesis on the emerging of the new ultra-conservative and populist forces in Brazil. Indeed, the lack of dialogue among research conducted from diverse perspectives—such as those of right-wing movement activists or voters—has hindered the attainment of a more comprehensive understanding of the various processes and dynamics associated with the reorganization of Brazilian society. This book aims to address these deficienciespresenting a diverse array of research that examines different aspects of the reconfiguration of the Brazilian right. The chapters encompass a wide range of topics, from the activities of right-wing movements both offline and online to the behaviors of voters aligning with the new right-wing factions. Additionally, the book addresses phenomena directly or indirectly linked to the ascent of the new right. In this context, the “new” right emerges, where the term “new” reflects the adoption of new strategies and mechanisms of political action, and not simply an ideological shift. The escalating polarization of the electorate, the reinforcement of anti-party sentiments, including anti-petismo and disdain for political parties in general, and the growing significance of identity issues and individual rights in electoral competition.
DISSERTATION
Dissertation Title: Experiments on Political Reasoning & Knowledge. Stony Brook University (May 2022).
Committee: Jennifer Jerit (chair), Stanley Feldman, John B. Ryan, and Efrén O. Pérez (external member).
SELECTED WORKING PAPERS (all available upon request)
Understanding Latino Political Engagement on Social Media (with Marisa Abrajano, Marianna Garcia, and CSMaP colleagues). forthcoming.
How Reliance on Spanish-Language Social Media Predicts Beliefs in False Political Narratives amongst Latinos (with Marisa Abrajano, Marianna Garcia, and CSMaP colleagues). forthcoming.
Issue Frames Under Multiple Information Enviroments (with Wladimir Gramacho and Max Stabile). Review & Resubmit.
Social media is usedmillions of Americans to access news and politics. Yet there are no studies, to date, examining whether these behaviors systematically vary for those whose political incorporation process is distinct from those in the majority. We fill this voidexamining how Latino online political activity compares to those of whites as well as the role of language in Latinos’ online political engagement. We hypothesize Latino online political activity to be comparable to white Americans. Moreover, given media reports suggesting that greater quantities of political misinformation are circulating on Spanish versus English-language social media, we expect that reliance on Spanish-language social media for news predicts beliefs in inaccurate political narratives. Our survey findings, which we believe to be the largest original survey of online political activity of Latinos and whites, reveals support for these expectations. Latino social media political activity, as measuredsharing/viewing news, talking about politics, and following politicians, is comparable to whites, both in self-reported and digital trace data. Latinos also turned to social media for news about COVID-19 than did whites. Finally, Latinos relying on Spanish-language social media usage for news predicts beliefs in election fraud in the 2020 U.S. Presidential election.
False political narratives are nearly inescapable on social media in the U.S. They are a particularly acute problem for Latinos, and especially for those who rely on Spanish-language social media for news and information. According to a recent studyNielsen (DeArmas, 2021), Latinos are vulnerable to misinformation, because they rely more heavily on social media and messaging platforms than non-Hispanic whites. Moreover, fact checking algorithms are not as robust in Spanish as they are in English, and social media platforms put far more effort into combating misinformation on English-language media than Spanish-language media, which compounds the likelihood of being exposed to misinformation. As a result, we expect that Latinos who use Spanish-language social media to be more likely to believe in false political narratives when compared to Latinos who primarily rely on English-language social media for news. To test this expectation, we fielded the largest online survey to date of social media usage and belief in political misinformation of Latinos. Our study, fielded in the months leading up to and following the 2022 midterm elections, examines a variety of false political narratives that were circulating in both Spanish and English on social media. We find that social media reliance for news predicts one’s belief in false political stories, and that Latinos who use Spanish-language social media have a higher probability of believing in false political narratives, compared to Latinos using English-language social media.
What are the effects of an issue frame before and after it gains salience in the real world? We offer an account of the long-term effects of issue frames over a 40-month period while keeping issue selection constant. We leverage the natural variation in issue salience in the information environment producedthe 2019 Brazilian Social Security Reform employing a combination of five survey experiments (N=8,277) and content analysis of news media (N=2,862). We supervise an automated content analysis of the Brazilian news articles to determine the issue salience of social security policy in the information environment. We demonstrate a substantial variation in issue framing effects before, during, and after the Social Security issue becomes a salient issue in Brazilian news media. Our findings show that novel information gained through experimentation triggers anticipated shifts in opinions when in low-information environments, but these effects dissipate when experiments are conducted within information-rich environments. Ultimately, if we had conducted each study in isolation, we would have drawn very different conclusions about public opinion and issue framing effects at each time.
Affective Homophily: A Motivated Reasoning Account for Why Persuasion Fails. Under review.
Theories of attitude change show that strong persuasive messages often times fail to persuade citizens. We develop a dual process network model based on the theory of political motivated reasoning to account for why persuasion so often fails. Here, political evaluation is driventhe hot cognition postulate which posits that all social-psychological concepts are affectively tagged. By this account it is not cognitive content, but affective homophily – the preferential of political associations – that drives the evaluation of a persuasive message. We use a thought-listing experiment to test how pre-existing positive or negative associations affects the evaluation of a persuasive message on the issue of genetically modified foods. From the resulting network of mental considerations, we test the following hypotheses: Hot Cognition → → Affective Homophily → Motivated Reasoning. First, we show that the content and structure of core associations is best characterizedaffectively congruent associations. Second, we test how persuasive messages change not only the cognitive content of survey responses but also the connectivity among core associations. Finally, we provide evidence that affective homophily it is what gives a political attitude its sense of coherence and the motivation to defend pre-existing beliefs against counter-attitudinal political information.
Yay or Nay? Brazilian Impeachment’s vote prediction from Political Speech. Under review.
In 2016, Dilma Rousseff, the 36th president of Brazil, was charged with criminal administrative misconduct and disregard for the federal budget in violation of the Constitution of Brazil and the Fiscal Responsibility Law. Here, I take the collection of vote speeches (504 valid votes out of 513 members of congress) during the Impeachment session in 2016 to predict ‘yay’ and ‘nay’ votes, and civil or uncivil rhetoric. Language is the channel for politics and an integral part of political conflict. Politicians constantly battle to secure the linguistic high ground, to control the language of a debate, and hence to frame the manner in which issues are considered. Through a split-sample design, I employ a deep averaging neural network (DAN) architecture inputted with word embeddings. The classifier maps the speeches into our measure of interest: the impeachment vote. While the model is syntactically ignorant, I show significant improvements over previous bag-of-words modelsdeepening the neural network and applying a novel variant of dropout leading to achieve a 90% accuracy in the speech’s classification. The high-level of accuracy achieved here shows that MCs were using clear-cut different words and language to vote in the 2016 impeachment.