My research examines how democratic institutions structure political power, representation, and political behavior. I work across voting rights and election law, democratic theory and measurement, and computational political behavior, with a particular interest in questions that require clearer links between political concepts, empirical evidence, and research design.
Across these areas, my work asks how political science can improve the way we identify, measure, and explain politically consequential forms of power. My current projects include research on the post-Callais landscape for race and redistricting, the role of political objectives and public disclosure in evaluating electoral opportunity, the measurement of shared political power in democracy, and the use of interpretable machine learning to recover structure in political behavior.
Voting Rights, Redistricting & Election Law
My current work in voting rights and election law focuses on the changing evidentiary landscape surrounding race and redistricting. I am particularly interested in how recent legal developments alter the kinds of empirical claims political scientists are asked to make, the evidence required to distinguish racial and political explanations, and the research designs available for evaluating those claims.
This work is increasingly concerned not only with particular redistricting disputes, but with a broader field-facing question: how should political science adapt when the evidentiary framework governing race, politics, and representation changes?
Political Science After Callais
Research agenda / field-facing paper in preparation
This paper examines the post-Callais landscape for political science research on race and redistricting. It focuses on how changes in the evidentiary framework reshape the empirical questions researchers must answer and the kinds of evidence required to evaluate claims about race, politics, and representation.
The paper develops a research agenda for the field, asking how political scientists should respond to a legal environment in which the relationship between racial and political explanations has become increasingly central to redistricting disputes. More broadly, it considers how changes in legal doctrine can alter the empirical paradigm through which political scientists study race and electoral institutions.
The Missing Benchmark: Political Objectives in Redistricting
Manuscript in preparation
This project examines a foundational but often underdeveloped problem in the empirical study of redistricting: identifying the political objectives against which districting outcomes should be evaluated.
The paper asks what states must disclose about the political and institutional objectives guiding redistricting if researchers, courts, and the public are to assess how those objectives shape electoral opportunity. Without a sufficiently specified benchmark, it is difficult to determine what observed districting outcomes should be compared against or to evaluate competing explanations for how a map was produced.
The project connects questions of research design and empirical identification to broader concerns about transparency, democratic accountability, and the exercise of state power over political representation.
Democracy, Shared Rule & Political Power
A second research program examines whether democracy can be adequately evaluated without considering how political power is distributed across the groups that constitute the political community.
This work focuses on shared rule: the degree to which groups participate meaningfully in governing power rather than simply possessing formal political rights or access to elections. It asks whether existing approaches to democratic measurement sufficiently capture the distribution of political power and what is lost when they do not.
Democracy and the Problem of Shared Political Power
Manuscript in preparation
This project argues that the extent to which political power is shared across groups is an important but insufficiently measured dimension of democracy.
The paper develops the Power Sharing Index (PSI), a group-centered measure designed to capture variation in shared political power. Rather than evaluating democracy only through the presence of institutions, rights, or electoral competition, PSI asks whether governing power is meaningfully distributed across the groups that make up the political community.
The project examines what the distribution of political power reveals about democratic inclusion and democratic quality. Its broader claim is that assessments of American democracy—and potentially of democratic quality more generally—remain incomplete when they do not account for the degree to which political power is actually shared.
Herrenvolk Democracy: Race, Exclusion, and the American Political Order
Book project
My book project examines how American democracy developed alongside durable racial inequalities in political membership and political power. It uses the concept of Herrenvolk democracy to study political orders in which democratic institutions coexist with racialized boundaries around membership, representation, and governing authority.
The project traces changes in political membership and power across major periods of American political development, with particular attention to the relationship between democratic inclusion and racial exclusion.
“Democracy for Whom? Herrenvolk Democracy and the American Founding”
Published in PS: Political Science & Politics, 2026 | Open access
This article examines the American founding through the concept of Herrenvolk democracy. It shows how the development of democratic institutions occurred alongside racialized boundaries around political membership, with particular attention to citizenship, popular sovereignty, and the Naturalization Act of 1790.
Open-access Article - Cambridge University Press
Political Behavior & Interpretable Machine Learning
My research on political behavior examines how electoral coalitions become internally differentiated and how the structure of political choice changes across elections.
A central methodological concern in this work is how machine-learning models can be used for more than prediction. I am interested in developing transparent and replicable strategies for using interpretable machine learning to identify which features structure political behavior, how those relationships vary across groups and elections, and how predictive models can be translated into substantively meaningful political evidence.
The Political Fault Lines in the Latino Electorate
Manuscript in preparation
This project examines political divergence within the Latino electorate across recent presidential elections.
Rather than treating changes in aggregate Latino vote share as the central outcome to be explained, the project asks what most strongly differentiates Latino voters from one another politically and how those fault lines change over time. Using interpretable machine learning, I examine the structure of presidential vote choice across the 2016, 2020, and 2024 elections and how the relative importance, direction, and magnitude of relevant political and social characteristics shift across electoral contexts.
The project uses these changing patterns to better understand partisan alignment, within-group heterogeneity, and the evolving structure of Latino electoral politics.
From Prediction to Structure in Political Behavior
Methodological paper in preparation
This project develops a replicable framework for using random forests and SHAP interpretation in political-behavior research.
The paper addresses a recurring problem in applications of machine learning to political science: predictive performance can identify models that classify outcomes well without, by itself, explaining the substantive structure underlying those predictions. The framework therefore focuses on how researchers can move from prediction to interpretation by using random forests and SHAP to examine the features that structure political choices, the direction and magnitude of their contributions, and variation across groups or contexts.
The broader goal is to provide a transparent and reusable workflow for scholars who want to use machine learning not simply to predict political outcomes, but to generate interpretable evidence about political behavior.
Related Work on Latino Politics & Measurement
“The Myth of the Zero-Sum: Rethinking Acculturation in American Politics”
Forthcoming, Politics, Groups, and Identities
This paper rethinks how acculturation is conceptualized and measured in political research. Rather than assuming that adopting American culture requires giving up one's heritage culture, I develop a multidimensional approach that captures several ways individuals can combine cultural adoption and retention.
Using the 2006 Latino National Survey, the analysis shows that many Latinos occupy cultural positions—including biculturalism—that conventional measures of acculturation overlook.
APSA Preprint | GitHub Code | Harvard Dataverse Replication Data
“Acculturation, Identity, and Latino Politics”
Under review
This paper examines whether different patterns of acculturation are associated with distinct political attitudes and identities among Latinos. It focuses on how cultural orientations relate to immigration, political belonging, identity, and other dimensions of political incorporation.
APSA Preprint | GitHub Code | Replication Data
Related public scholarship: “Latino Support for Trump in 2024: Trends and Insights from an Empirical Analysis,” with Luis R. Fraga, Political Science Now.