Beyond the Structure of Attitudes: Belief Systems, Affect, and Polarization
Despite the recognition of the importance of affect elsewhere (including in much work on polarization), work on public opinion and belief systems has so far looked only at propositional attitudes. But, as seen in the worldview model of polarization, the connections between propositional attitudes and affective judgments are of central importance to the actual function of any given belief system (or worldview). Through a novel application of Boutyline and Vaisey’s belief network analysis, this paper demonstrate this importance in a new way. Using data from the American National Election Studies, I compare the belief networks in 1992 and 2016. But beyond examining only the network of propositional beliefs, I also look at the network of affective items, as well as the combined network of both attitudes and affects. In this way I show, first, that the different kinds of networks are similar enough to be comparable, but different enough to provide analytical leverage; and second, that the integration of affects and beliefs has changed over time, in line with polarization. Ultimately, I argue that we must incorporate affects into our study of belief systems and, furthermore, that the nature of this incorporation is itself a meaningful axis of variation. These findings are not only important methodologically for the study of belief structures, they are substantively consequential for the study of polarization, offering a way towards reconciling affective and attitudinal approaches by defining polarization in terms of the increasing integration of emotion and attitude.
Draft available upon request. PowerPoint slides and notes from presentation at ASA in 2023 are available here.
Talents for the Talented: Disrupting the Matthew Effect in an NSF Award Program
The Matthew effect (or cumulative advantage) in citations distorts the connection between merit and recognition in science, hindering the careers of individual researchers as well as the progress of science itself. In this paper, we assess possible ways of disrupting the Matthew effect in a system of knowledge production by studying the NSF’s ADVANCE program. Specifically, we examine the program’s efforts to promote the integration of social science research on gender equity in their funded projects. By tracing changes in ADVANCE’s award solicitations over time, specifically the citation of certain works and the introduction of a mandate to address intersectionality in all proposals, we evaluate their impact on the citation distribution in the resulting publications from ADVANCE awards. Using a combination of timing-based evaluations and recent advances in generative network modeling, we establish quasi-experimental comparisons between treated and untreated sources, as well as pre-treatment and post-treatment sources. Our findings suggest that both interventions increased citations to authors highlighted by the changes, mitigating the Matthew effect and enhancing the role of source quality in driving citations for at least some sources. Notably, the intersectionality requirement had more consistent and robust effects. These findings both provide a rare example of successful agentic intervention in a complex system as well as indicate a potential difference in the efficacy of targeted and criterion-based interventions.
In collaboration with Laura Nelson, Jessica Gold, and Kathrin Zippel.
Draft available upon request. Replication materials for this paper are available here and replication materials for the wider project will be available here.
Inactive and quiescent? Immigrant Collective Action in Comparative Perspective, 1960-1995
Social movement mobilization by and on behalf of immigrants occurs frequently today, but sociologists have been slow to include immigrant collective action in the canon of social movement or immigration scholarship. Is this because, until recently, immigrant protest was minimal or limited in scope? We take a macro-comparative approach, re-coding the Dynamics of Collective Action dataset to compare pro-immigrant collective action with paradigmatic, well-studied movements from 1960 to 1995. We find that immigrant collective action was on par with iconic movements, mobilized similar numbers of people, occurred across the United States, involved disruptive action, and encompassed a wide range of national origins, thus correcting possible misperceptions that immigrants did not engage in contentious action before the 1990s. We advocate for a population at risk focus for studying the emergence of collective action, for decentering the borders of collective mobilization, and for considering the vulnerabilities of legal status.
In collaboration with Kim Voss and Irene Bloemraad.
Forthcoming in Socius
What Makes a Movement? Domains of Contention and the Pre-History of the Pro-Immigrant Social Movement, 1960-1995
Building on our work in “Inactive and quiescent?”, this paper asks why the raw quantity of activism we identified in our recoding of the DoCA dataset did not coalesce into a sustained immigrants rights movement. To explain why the nearly seven hundred events did not snowball into a sustained movement we inductively identify four different domains of contention: diasporic, documentation, access, and exclusionary protest. The domains are differentiated not just by the key issue at stake, but also by different temporal trends in collective action, number of events, targets, groups involved, SMO engagement, and tactics. These differences mean that domain-bridging is difficult, undermining cohesion and sustainability over time. Furthermore, they also show different relationships between mobilization and threat, with repressive threat on the part of the government, not policy threat, mobilizing people reactively in the documentation domain, while the access domain more frequently saw proactive mobilization in response to opportunities or policy threats. The concept of domains gives us a new way to see patterns in inconsistent or episodic action and therefore a new way of studying collective action in the absence of a cohesive movement. More than that, we demonstrate how it can help us better investigate the precursors of such a movement, by illuminating the pre-history of the modern pro-immigrant movement.
In collaboration with Kim Voss and Irene Bloemraad.
Draft available upon request.
Polar Opposites: Real and Imagined Differences in Partisan Candidate Evaluation
Political scholars are grappling with the causes and consequences of political polarization in U.S. politics. One area of great concern is related to voting behavior, especially in an era of uncivil politics. As candidates eschew norms of decorum for uncivil tactics (e.g. insulting opponents or spreading misinformation), voters find themselves electing uncivil candidates and long-held beliefs about voters’ motivations are called into question. What remains unclear from existing literature is how (1) how Americans justify choosing an uncivil in-party candidate; (2) what information Americans will leverage when forced to choose between two in-party candidates that vary in their uncivil behavior; and (3) how Americans judge civil out-party candidates; and (4) how do respondents leverage candidate race and gender in their justifications? This research measures systematic differences in the way Americans evaluate civil and uncivil in-party and out-party candidates. This study puts into conversation three important trends in political sociological scholarship: (a) the identification of voter motivations; (b) the foundational question of in-party affect versus out-party animus; (c) voters’ response to candidate incivility. We do so by applying automated text analysis techniques to more than 10,000 survey write-in responses from 1,737 U.S. adults. In our preliminary results, we identify important themes from the responses. First, partisan alignment is the deciding factor for most respondents. Second, when partisan alignment cannot be the deciding factor (when respondents are choosing between two in-party or out-party candidates), respondents tend to weigh the relative incivility of the candidates as well as perceived candidate quality based on experience and candidate demographics. Our findings suggest that voters are choosing uncivil candidates with open eyes – recognizing that partisan alignment is more important than candidates’ behavior during campaigns.
In collaboration with Jennifer Dudley.
Draft available upon request.
Privacy in Public? The Ethics of Academic Research with Publicly Available Social Media Data
The public nature of social media data has been an incredible boon to researchers across disciplines, providing access to a previously impossible quantity and (in certain respects) quality of data. Part of the attraction of this data is that it has traditionally traditionally been understood as exempt from the ethical and regulatory responsibilities applied to purpose-collected or otherwise private data. However, the very attributes which make this data uniquely useful also raise unique ethical challenges. This paper attempts to explore those challenges and their implications for researchers by working through the differences between “public on the internet” and “public in physical space.”
We focus in particular on what these differences mean for the social media users we study and ground our evaluation of harms in the perspective of researchers’ “duty of care” towards those users whose data we collect. We empirically explore the conflict between users’ interest in controlling their online self-presentation and researchers’ interest in maintaining a durable record by checking the continued availability of previously collected tweets. We find a large proportion of deleted and private tweets, demonstrating a relatively common desire to opt-out of future, and presumably therefore previous, study. In the second part of the paper, we discuss the institutional and regulatory frameworks governing the use of such data. These are the mechanisms through which any possible solutions must function, but they are not as yet suited to the task. We close with recommendations for moving forward, attempting to strike a reasonable balance between ethical ideals and scientific realities.
In collaboration with Nataliya Nedzhvetskaya.
Available in the 2023 issue of the Berkeley Journal of Sociology or here.
Strangers in their Own Party: Dissatisfied Partisans and the Worldview Model of Polarization
The Republican Party’s rejection of the outcome of the 2020 presidential election is nearly unprecedented in modern American politics, but there is one way that it is unremarkable: it has no Democratic equivalent. Rather, it is only the most extreme example of the apparently asymmetric nature of polarization over the last forty years. However, current work on mass polarization neither captures nor explains this asymmetry. In this paper, I bring together sociological work on cultural schemas, recent advances in the study of attitudinal polarization, and core insights on the role of affective polarization from political science, to propose a new theory which can and does: the worldview model of polarization. By defining polarization in terms of persistent patterned relationships between partisan affect, perceptions of threat, and feelings of fear, I am able to approach asymmetry as a question not only of degree, but of kind. Using data from the 1978-2016 ANES, I not only clearly demonstrate that Republicans are indeed significantly more polarized at the mass level, but also suggest a mechanism for this asymmetry, namely, that Republicans are caught in a radicalizing feedback loop of fear of the opposition and dissatisfaction with their own party’s response to that perceived threat.
Negative Partisanship is Not Enough: Affective Polarization, Partisan Asymmetry, and a Failed Theory Contest
Despite the increasingly obvious centrality of affective polarization to American politics, a foundational question remains unresolved: from the start, some work has defined polarization solely in terms of out-party animus, while others have seen both in- and out-party affect as constitutive. In this paper, I test these approaches on the question of whether mass polarization is asymmetric: given evidence at the elite and organizational levels, we expect Republicans to be more polarized than Democrats, and these two definitions create clearly different expectations for how that asymmetry should look. However, this apparently neat theory contest fails. Using data from the American National Election Studies, I look at how affective polarization developed from 1978–2016 and find asymmetry toward Democrats, driven almost entirely by in-party affect. These findings are anomalous for both theories, and I argue that they disrupt our basic assumption of a straightforward link between affect, identity, and polarization.
Draft available upon request.
Solidarity and Strife: Democracies in a Time of Pandemic
In 2020, I was made a research fellow with the Solidarity and Strife project, a collaboration between the Social Science Matrix and the D-Lab which aimed to use social media data to take “a comparative approach to assessing the impact of the COVID-19 pandemic on political polarization and solidarity in the United States and the United Kingdom, two nations governed by populist leaders who initially denied the seriousness of the viral outbreak.” My research from this project is still ongoing, but it attempts to use an extensive Twitter dataset to investigate exactly how polarization began and developed in conversations around COVID-19.
Speaking Sociology: The Problem of Polysemy and the Solution of Metaphor
Doing good sociology is hard, this goes without saying. But the specifically semantic or communicative dimension of this difficulty is often ignored. We sometimes imagine sociology as a diverse discipline in which crosscutting perspectives and methods collide in strange and exciting conversations, but in reality an outsider would be forgiven for thinking that sociology only purports to be a single discipline, as these semantic difficulties too often create misunderstandings and see our conversations siloed off. Building on Abend’s (2008) recognition of the problem of polysemy in theory, I show that the problem is far more general, impacting every corner of the discipline. Previous solutions, while helpful, have been insufficient because they lack an adequate theoretical understanding of communication. However, the philosophical, linguistic, and cognitive study of metaphor can provide just such a foundation. Building on this foundation, I detail a practical solution, which I dub metaphorical reading, and I demonstrate its utility with a series of practical examples.
Draft available upon request.