Published Works
Gervais, Bryan, Connor Dye, Gabriel Acevedo, Christopher Ellison, Margaret Kelly. “Do Authoritarians Support Political Violence?” PS: Political Science & Politics.
Gervais, Bryan, Connor Dye, and Amebr Chin. 2025. “Incivility or Invalidity? Evaluating Perspective API Scores as a Measure of Political Incivility.” American Politics Research 53(3): 266-274.
Shafran, JoBeth, Bryan Jones, and Connor Dye. “Bounded Rationality and Public Administration.” In The Oxford Research Encyclopedia of Politics. Oxford: Oxford University Press.
Working Papers
“Overcoming Information Overload: The Policy Consequences of Committee Fragmentation” Revise and Resubmit
Does overlapping committee jurisdictions expand or constrain Congress's capacity to understand complex policy problems? Despite extensive theoretical debate, scholars disagree on whether multiple committees examining the same issue increase informational diversity or merely generate competing narratives from identical evidence. I analyze 92,384 instances of agency testimony across 131 agencies from 1973 to 2021 to assess how committee structure affects the diversity of bureaucratic information incorporated into congressional decision-making. Examining variation in both the breadth and concentration of agencies testifying across issue areas and time reveals that more committees are consistently associated with a more diverse set of agencies supplying information on problems. Subsequent results demonstrate that the relationship between informational diversity and the number of committees examining an issue operates identically during periods of policy stability and punctuation, indicating that committees rely on distinct bureaucratic networks rather than altering their information search in response to crisis. Decentralization therefore improves Congress’s capacity for problem definition even as overlapping jurisdictions may complicate downstream control of the bureaucracy.“Academic Utilization and Issue Novelty” Under review
Why do policymakers utilize academic expertise when alternative sources provide more accessible information? I argue that academia's evaluation system decouples research priorities from policymaker demand, enabling academics to accumulate expertise on emerging problems that utilization-oriented organizations systematically avoid until issues become politically salient. I test this theory examining expert witness selection at U.S. congressional hearings from 1969 to 2021. Using natural language processing methods to measure issue novelty through hearing descriptions, I find that academics constitute a significantly larger share of witnesses when issues are novel and a smaller share as issues mature. Think tanks, federal agencies, and analytical agencies show the opposite pattern, providing evidence that increased academic testimony on novel issues reflects compositional shifts rather than broader demand for expertise. These findings suggest that academic utilization stems from its unique evaluation system that creates incentives for academics to invest in problems that other experts avoid, and that efforts to make academic research more policy-relevant may undermine the distinctive value academics provide when policymakers confront unfamiliar issues.“The Erosion of Specialization: How Institutional Evolution Undermines Congressional Oversight” Under review
Congress’s decentralized committee system expands oversight capacity but creates coordination problems. I argue that the incentives behind coordination challenges evolve as the committee system responds to emerging problems. Initially, informal domain specific expertise accumulated through legislating mitigates coordination challenges by lowering oversight costs in familiar areas. As boundary-spanning issues emerge, committees' legislative jurisdictions converge, eroding expertise differentials and increasing duplicative oversight. To capture jurisdictional evolution, I introduce a novel computational approach using transformer-based embeddings of House bill summaries and cosine similarity to track both within-committee jurisdiction change and cross-committee convergence. I find that committees with larger jurisdictional shifts subsequently oversee more novel agencies, while committee pairs with greater prior convergence show higher probabilities of duplicating oversight on identical agency-subtopic combinations. These findings suggest that effective oversight design must account for how coordination mechanisms deteriorate as policy environments evolve.Works in Progress
“Information Processing and Congressional Reform” (with Sean Theriault)
This paper reevaluates the Legislative Reorganization Act of 1946. While the Act’s contemporaries may have been underwhelmed, we show that the jurisdictional changes it mandated had a significant consequence on congressional attention. At a broad level, the committee system operated differently after the Act’s passage. A more fine-grained analysis shows even more compelling evidence. Committees that experienced jurisdictional changes had relatively more hearings; furthermore, issues that underwent change were the subject of more congressional attention.“Constrained Campaigns: How Alliances Reduce Party Attention to Defense Issues” (with Kevin Galambos)
How do international commitments influence domestic political party competition? While scholars have extensively examined how international affairs affect domestic politics through leaders and elites, the effect of international commitments on political party dynamics remain underexplored. In this article, we argue that alliance commitments constrain parties' abilities to credibly differentiate themselves on defense issues by creating domestic and international audience costs for deviating from supporting security partners. Given these constraints, parties reduce their emphasis on defense issues and exhibit less variation in their foreign policy positions. Using data from the Comparative Manifesto Project covering 41 countries between 1991 and 2023, this study shows that NATO and Partnership for Peace membership is associated with decreased emphasis on defense issues during election campaigns. Subsequent results demonstrate that this lack of attention is driven by convergence, as parties in member countries exhibit significantly less variation in defense positions compared to non-member countries.“Geographic Framing and Immigration Attitudes: How Local Versus National Perspectives Shape Public Opinion” (with Megan Dias)
Public preferences toward immigration policy are shaped not only by perceptions of immigrants but also by the geographic scale at which immigration is considered. While prior research emphasizes economic and cultural threat as drivers of immigration attitudes, we argue that geographic framing shapes public opinion by altering how people psychologically process immigration-related information. Drawing on construal level theory, local frames prompt concrete thinking about specific immigrants in one's community, activating inclusive community identities and more favorable attitudes. In contrast, national frames promote abstract thinking about immigration as a phenomenon, activating exclusive national identities and less favorable attitudes. To test this theory, we plan to field a nationally representative survey experiment in which respondents are randomly assigned to receive either a local, national, or neutral immigration frame. We then plan to measure attitudes toward both local immigrant integration policies and national-level enforcement policies. We will further examine whether local immigrant concentration moderates the effect of geographic framing, by linking survey respondents to local demographic data..“Ballot Initiatives, Question Wording, and Voter Intent: Assessing Voter Understanding of Ballot Propositions during the 2024 Election “ (with Bryan Gervais, Elias Hudson, Camilo Nieto-Matiz, and Walter Wilson)
While democratic theory has espoused the virtues of ballot measures, political scientists have questioned whether ballot propositions empower voters or organized, established interests. We contribute to this debate by considering how voters support for several propositions significantly changed, depending on whether they were presented with common descriptions of the measures featured in local news media or the actual, vague ballot wording. Our study demonstrates that the language used to present ballot initiatives can profoundly shape public support. While simplified descriptions drawn from local media coverage initially elicited low levels of support in an initial survey, the actual ballot language featured in a second survey—despite its more technical and legalistic phrasing—produced significantly higher approval. In a third study, featuring respondents from the first two surveys, we employed an experimental design, which leveraged both between-subject and within-subject comparisons across waves, The results confirmed the robustness of this effect: participants frequently changed their positions depending on whether they encountered a plain-language summary or the official ballot text. These shifts were not random but systematically driven by the form of wording.“Party Identification and the Subjective and Objective Economic Vote” (with Christopher Wlezien, Mark Kayser, Bailey Via, and Jan Zilinsky)
Numerous studies have demonstrated a weakening identification of voters with political parties in Western Europe over the last 40 years. Previous research shows that this has led to increased economic voting, where less partisan voters place greater weight on (objective) macroeconomic conditions. Building on research demonstrating that individual-level economic perceptions are in part endogenous to party preference, we argue the degree to which economic perceptions are brought in line with vote choice also depends on partisan attachment. Specifically, we expect that trends in partisanship produce contrasting tendencies in the effects of the objective and subjective economy on the economic vote: low attachment reduces the subjective economic vote while increasing the extent to which objective economic conditions are considered in the voting calculus. The model not only accounts for trends in Europe, where partisanship has declined, but it also helps explain voter behavior in places where identification with parties has increased.