Normally Political
The declarations of independents: Open-ended survey responses and the nature of non-identification
Maxwell Allamong et al.
American Journal of Political Science, forthcoming
Abstract:
While many Americans identify as politically "independent," conventional wisdom suggests most are covert partisans, especially "leaners." However, we argue that independents exhibit distinct attitudes toward political parties. Analyzing American National Election Studies open-ended responses from 1984 to 2020, we employ structural topic models and support vector machines to explore differences in how independents and partisans express "dislikes" about the major parties. Our results show that leaning independents differ significantly from self-identified partisans in their language and align more closely with pure independents. Furthermore, we find that independents are more likely than partisans to mention issues related to "politics" -- such as political cynicism, concern about special interests, or distaste for partisan conflict -- when discussing their "dislikes." These findings suggest that independents have meaningful distinctions from partisans beyond voting behavior, cautioning against overlooking the nuances between these groups.
Ideological Moderation and Success in U.S. Elections, 2020-2022
Michael Bailey & Benjamin Reese
Journal of Politics, forthcoming
Abstract:
Understanding how candidate ideology affects elections is crucial to understanding representation. We assess the influence of ideological positioning on election outcomes for virtually all general election candidates in the 2020 and 2022 U.S. House, Senate and gubernatorial elections. Our approach has three novel features. First, we use original ideology estimates based on the language candidates use on social media and their campaign websites. Second, we leverage the fact that multiple elections occur simultaneously, allowing us to compare the performance of ideologically distinctive candidates within a district. If a moderate candidate in one race outperforms an extreme candidate in another race among the same electorate, we have evidence that moderation is beneficial. Third, we compare patterns across federal, state, legislative and executive positions. Across multiple specifications and multiple comparisons, we find that ideological moderation was associated with higher vote shares, especially in competitive and gubernatorial races.
Reassessing Extremism, Polarization, and Constraint with Continuous Policy Questions
Anthony Fowler
Political Behavior, forthcoming
Abstract:
Some argue that the American public is extreme and polarized along party lines. Paradoxically, others argue that members of the public lack meaningful policy preferences and exhibit low constraint across issues. These conclusions are typically drawn from binary policy questions or scales with ambiguous values, both of which are ill-suited for measuring extremism, polarization, or constraint. In this paper, I reassess these claims by analyzing policy questions that allow respondents to express their preferences on a well-defined continuum. Across a wide range of issues, most Americans appear to have moderate preferences over policy. As expected, Democrats tend to be more liberal than Republicans, but there is significant overlap on every issue, and the average extent of disagreement is modest. Lastly, positions across issues appear more constrained than standard tests suggest.
Reassessing Support for Political Aggression and Violence in the United States
Scott Clifford, Lucia Lopez & Lucas Lothamer
Public Opinion Quarterly, forthcoming
Abstract:
Recent events have driven a surge in scholarly attention to public support for political violence in the United States. Yet, research paints a conflicting picture about the levels and correlates of support for violence. We argue that these disagreements are partly due to researchers' measurement choices. After reviewing common practices and identifying measurement challenges, we introduce a measure designed to overcome these problems that allows respondents to choose their target of aggression. Across multiple studies, we compare our measure to two common alternatives. While we find similarities, our measure uncovers substantially more support for aggression and violence, particularly among weak partisans, holding implications for the levels and correlates of support for aggression. Further, by design, our measure provides information about the type of aggression that is endorsed and the most common targets. We conclude with recommendations for researchers studying support for political aggression.
Unaggrandizing the Partisan Gerrymander
Michael Barber, Andrew Taylor & Sean Trende
Election Law Journal, forthcoming
Abstract:
Observers suggest the effects of redistricting with excessive partisan bias or gerrymandering are particularly distinct and durable today. In this telling, partisan polarization, sophisticated technology, and detailed information of voter behavior permit those who draw state legislative and congressional plans to secure favorable electoral outcomes for their party throughout the decade the plan is in effect. We provide an empirical test of this basic claim by comparing three plan "pairs" -- essentially juxtaposing plans accused of exhibiting excessive partisan bias by academics, journalists, or the courts with equivalents that are not. Using traditional indicators of partisan bias in redistricting, we find evidence that gerrymandered plans differ less than implied by characterizations from the non-gerrymandered plans with which they are paired and that any advantage given to a party at plan conception tends to decline over the course of a decade as partisan bias in the gerrymandered maps declines and differences in plan pairs alleviate. We conclude with thoughts regarding the implications of our results for political scientists, potential litigants, and jurists.
Assessing Ideological Representation Using Policy Repositioning: Panel Evidence From 2012 to 2020
Andrew Gooch
American Politics Research, forthcoming
Abstract:
Assessing ideological representation can be difficult because perceived ideology and party identification correlate strongly among presidential nominees. Democratic nominees are consistently liberal, and Republican nominees are consistently conservative. However, Trump's 2016 candidacy included ambiguous, liberal, and extremely conservative positions. While in-office, Trump repositioned to uniformly conservative policy positions. This presents an opportunity to assess changes in candidate ideology that are difficult to observe outside of experiments. With unique panel data, I used self-placed ideology in 2012 to create spatial distances to Trump and the Democratic nominees in 2016 and 2020. This design allowed for over-time variation of candidate evaluations while keeping self-placed ideology pre-Trump. I find that Trump was perceived as moderate in 2016, and this translated into a vote choice advantage relative to Clinton. However, respondents substantially reevaluated Trump as being conservative in 2020, which changed the vote choice advantage to Biden. These results show that perceived ideology influenced voting even in an era of strong partisanship, and this typically cannot be observed because of the ubiquitous connection between ideology and party among presidential candidates.
Abortion policy preferences are structured, stable, and consequential
Natalie Hernandez et al.
American Journal of Political Science, forthcoming
Abstract:
Do Americans have structured, stable, and consequential policy preferences that shape political outcomes? We explore this question through the case of abortion, using a large-scale panel dataset (n ? 130,000) and applying three key diagnostics: coherence, stability, and changes in vote choice. First, we demonstrate that abortion policy preferences exhibit logical coherence, both within and across reasons for seeking an abortion. Second, we show that these preferences are highly stable over time -- more so than personality traits -- suggesting that abortion attitudes are deeply engrained rather than fleeting opinions. Lastly, we find that abortion policy preferences, measured before the U.S. Supreme Court overturned Roe v. Wade in 2022, predict shifts in intended voting behavior between 2020 and 2024. This overall pattern helps rule out key theoretical alternatives, such as non-opinions, attitudes following vote choice, and elite cues. Additionally, these findings highlight the significant and independent role of abortion attitudes in shaping American political behavior.
News of Converging Crises: Cable Coverage of Race and COVID Stories
Brian Robert Calfano, Valerie Martinez-Ebers & Aida Ramusovic
Political Research Quarterly, forthcoming
Abstract:
Drawing on theories covering media differentiation, delegitimating, and journalistic professionalism, we show FOX is substantially more likely than CNN or MSNBC to increase its coverage frequency of race stories in response to topical coverage by its competitors. CNN reduces its coverage frequency of these race stories in response to FOX (and MSNBC follows this same general pattern). By contrast, all three networks reinforce each other's COVID coverage. The implication from these findings is that even media outlets with known ideological brands do not agenda set independently of the wider media ecosystem in which they operate.
The Good, the Bad, and the Not-So-Terrible: A Comprehensive Analysis of Sinclair's Effect on Local Television News Quality
Jesse Abdenour et al.
Journalism & Mass Communication Quarterly, forthcoming
Abstract:
Critics accuse the Sinclair television ownership group of slashing resources and producing biased news. However, few scholars have examined Sinclair's news quality. Using an investment-based approach to quality, this study provides the first systematic human-coded examination of Sinclair's local video news quality. Data from 2,300 stories in 40 U.S. markets indicated lower overall quality at Sinclair-owned stations compared to competitors. However, Sinclair outperformed competitors on some quality metrics, and evidence of political bias was minimal. Findings provide insights into how corporate ownership affects local TV, the most trusted source of news in the United States.
Multidimensional Signaling and the Rise of Cultural Politics
Daron Acemoglu, Georgy Egorov & Konstantin Sonin
NBER Working Paper, February 2026
Abstract:
In turbulent times, political labels become increasingly uninformative about politicians' true policy preferences or their ability to withstand the influence of special interest groups. We offer a model in which politicians use campaign rhetoric to signal their political preferences in multiple dimensions. In equilibrium, the less popular types try to pool with the more popular ones, whereas the more popular types seek to separate themselves. The ability of voters to process information shapes politicians' campaign rhetoric. If the signals on the cultural dimension are more precise, politicians signal more there, even if the economy is more important to voters. The unpopular type benefits from increased conformity, which bridges the candidates' rhetoric and makes it more difficult for voters to make an informed decision.