Americans are exposed to varying levels of partisan hostility depending on where they live, and this exposure is likely to have political, social, and psychological ramifications. This study encourages a new focus on community polarization by presenting an original dataset that estimates polarization among the public, and among members of each party, in over 6,600 US cities and towns. The data show that polarization varies widely across communities. While some are defined by searing political animosity, in others, residents exhibit more “live and let live” attitudes. A mixed-method analysis of this variation shows that exurban communities are more polarized than others, that a party’s local strength corresponds with more polarization among its members, and that civic institutions have conditional effects on polarization: exacerbating it in heavily Democratic areas, but quelling tensions in more Republican communities. Presented as the first step in a research agenda on community polarization, the paper concludes by laying out paths for future study.

Researchers have historically thought of protesters as “outsiders looking in” at the political system, but the last two decades have highlighted the interconnectedness between protest movements and partisan politics. Parties recruit candidates from the ranks of movement activists, and strong partisan identifiers are more likely to protest than others. In an era defined by spikes in both partisan polarization and protest activity, this paper theorizes two ways in which the trends may relate. In the first, partisan animosity may serve as a form of “diffuse” threat — as opposed to the narrow threat of a policy action — that mobilizes people to protest. In the second, tolerance from the political opposition may indicate a more open cultural opportunity structure, increasing the perceived upside of protest. My findings support the second of these, but only among left-wing protest. An analysis of 19,170 protest events by US county in 2021 and 2022, as indexed by Google News, shows that counties with more tolerant Republicans see greater left-wing protest activity, and this holds when controlled for by state, presidential vote, and attitudes among Democrats.

A growing body of research explores the influence of interviewer race on the survey responses of Black Americans. In one strand, Black respondents are described as minimizing progressive viewpoints in the presence of White interviewers (e.g., Davis 1997). In another, Black respondents are understood to emphasize such viewpoints to Black interviewers (White and Laird 2020, Wamble et al. 2021). These interpretations are important not just methodologically, but substantively, as they speak to the contextual determinants of Black partisanship. We integrate these divergent narratives by leveraging variation in interviewer phenotype (using measures of race, ethnicity, and skin tone) to run two quasi-experiments. The results show both “play it up” and “play it down” effects to be at work, with much of the movement occurring among Black conservatives. Our findings validate and add nuance to previous research, and open questions about the future of Black partisanship in an increasingly diverse society.

Research in the US and Europe has found that the adoption of LGBT rights can reduce youth suicide, as tragically, as many as one-quarter of youth suicides occur among gender and sexual minorities. However, it is less clear whether this relationship would hold in the developing world, where the range of protections for LGBT citizens—as well as of cultural expressions of gender and sexuality—varies more widely. I use linear panel models with country-based fixed effects to assess the relationship between LGBT rights, LGBT INGOs, and the suicide rate for male adolescents in 93 non-OECD countries between 1990 and 2016. Contrary to expectation, no relationship is found between the diffusion of LGBT rights and youth suicide, but a separate relationship offers hope: one standard deviation of greater integration into global LGBT INGO networks associates with a 8.9% drop in male adolescent suicide relative to the 2016 median.

Protest activity in the US has spiked over the past 15 years, and social justice activists perceive unequal treatment by police at these events. Existing quantitative research supports these claims, but most studies rely on data from 20th century events, and few suggest policy remedies. Our project updates past findings using data on police arrests (or the absence thereof) at 110,000 protests from 2017 to 2022. Then, we identify factors and policies that can lead to fairer treatment from data on 2,800 local police departments. The data show that protest arrests are infrequent and situational, but that charges of unequal treatment are justified. The odds of protest arrests increase by roughly 80 percent when protesters demand police reform or gather in support of marginalized groups. However, these disparities can be mitigated. Racially diverse police forces, and oversight measures such as body cameras and civilian complaint review boards, associate with more equal treatment between social justice protesters and others. These findings update the literature on protest policing to the present era, and suggest directions by which policymakers can achieve public safety without neglecting citizens’ rights to free assembly and expression.

* This represents my contribution to a larger project on protest policing, conducted alongside researchers at the RAND Corporation, and funded by the National Institute of Justice.

Measuring ideology is essential to understanding why political actors behave the way they do. But creating measurements that allow comparisons between different types of actors—such as legislators and judges, or executives and voters—has been a vexing challenge for researchers. Bonica’s (2014) campaign finance-based ideology scores, or CFscores, respond to this challenge by drawing on patterns of “who donates to whom” to estimate the ideologies of donors and candidates for all types of offices on the same scale. However, most citizens are excluded from this common space because they do not donate to candidates. In this paper, I use a random forest machine learning model to impute CFscores for non-donating citizens based on how their survey responses compare to those of donors. These imputed CFscores outperform other measures of ideology in terms of explaining vote choice, and are robust to differences in issue importance between donors and non-donors. While the use of imputed CFscores for citizens is currently constrained by data availability, the results presented here should encourage researchers to explore and expand the uses of this measure.

Executives are important actors in the political system, and ideology plays an important role in elite behavior, but measurement challenges and a focus on the presidency have kept scholars from fully exploring the role of elite ideology. This article proposes that students of executive politics may be well-served by turning to the gubernatorial context, where campaign finance-based ideology scores (Bonica 2014) provide a common-scale measure of executive preferences. These scores, or CFscores for short, are drawn from 103 million campaign contributions based on patterns of “who gives to whom?” I demonstrate that CFscores converge with other indicators of executive preferences, and reliably predict executive behavior. Then, I examine the influence of executive ideology over policy. Four models explain state policy liberalism as a function of executive, legislative, and citizen ideology, and gubernatorial preferences emerge as the most predictive of the three. Executive ideology is more explanatory than legislative preferences when Democrats are in office and dwarfs the predictive power of public opinion in all cases. A one standard deviation shift in executive ideology corresponds with 2.3 to 8.3 times more policy change than a similar shift in public preferences.

Scandals are important events in democratic life, holding important implications for public trust, opinion, and policymaking. But how do scandals come to dominate the political agenda? I examine four Obama Administration scandals, using the daily tweet volume (Barberá et al. 2019) of three sets of actors — the public, elites, and media — to examine who drove whose attention to these incidents. The short- and long-run results suggest important differences. The public’s attention had a strong, instantaneous effect on elite tweet volume. But, when measured over time, elites had a greater impact on public attention than vice-versa. This presents the possibility of a “sounding board” process, by which elites test the resonance of news events, so as to focus their attention on those that will elicit the strongest reactions from their supporters.

Even prior to the spread of the novel coronavirus, vaccine uptake in the United States fell short of levels that achieve herd immunity. Efforts to boost vaccination rates have taken three tracks: government policies that require vaccination, incentives that reward vaccination, and health communication efforts that emphasize persuasion. Focusing on persuasion, we call attention to America’s army of health education teachers as an overlooked resource in pro-vaccination campaigns. Our analysis of state educational content standards – each state’s written policy concerning what students are expected to learn – shows that no states make vaccine literacy a high priority. Then, in turning to teachers, we use a nationally representative survey of health education teachers to estimate the attention that teachers give to vaccines. Nearly three-fifths of teachers never discuss vaccines with their students, but those who do overwhelmingly emphasize the scientific consensus around vaccines, and no political, religious, or racial differences exist between these groups. This sets the expectation that, if called upon, health education teachers would be willing and capable allies in the effort to boost vaccination rates.