Research
Research in American Politics and Political Methodology
Selected Working Papers
Political scientists often use aggregate spatial units like neighborhoods, precincts, and administrative districts. Statistics based on aggregate data are sensitive to boundary definitions, a well-known issue called the Modifiable Areal Unit Problem. Scholars struggle to investigate this issue because the number of possible aggregations in most applied problems is intractably large. We propose a sampling-based method for measuring geographic aggregation sensitivity. Building on graph partition algorithms developed for legislative redistricting, our approach allows researchers to sample alternative aggregations from a specified target distribution. Researchers can ensure aggregations are realistic for their setting by defining constraints, such as population targets and respect for geographic boundaries. Researchers can then compare estimates across aggregations. We illustrate our approach through applications measuring segregation, racially polarized voting, and the relationship between racial context and exclusionary attitudes. Alternative aggregations can shift estimates in these analyses by 11% to 44% depending on outcome and setting.
@misc{brown2026sampling,
title = {Sampling Solutions to the Modifiable Areal Unit Problem},
url = {osf.io/preprints/socarxiv/gpq32_v1},
publisher = {SocArXiv},
author = {Brown, Jacob and Kenny, Christopher T and Simko, Tyler},
year = {2026},
month = {Jul}
}
In April 2026, the US Supreme Court issued the Louisiana v. Callais decision, weakening the Voting Rights Act (VRA). We estimate the impact of the Callais decisionon minority representation and electoral competition in the US House under two scenarios: (1) congressional district boundaries are drawn in a race-blind, nonpartisan manner without complying with the pre-Callais VRA requirements and (2) both parties engage in partisan gerrymandering to maximize their seat shares without considering former VRA protections. Our analysis uses simulation algorithms to generate alternative redistricting plans under these scenarios while satisfying traditional redistricting principles and state-specific criteria. First, we show that in aggregate the pre-Callais interpretation of the VRA created similar levels of minority representation to race-blind nonpartisan redistricting, suggesting that the VRA did not create large partisan advantages. Next, we show that if states continue to aggressively gerrymander, the Callais decision is likely to benefit the Republican Party and reduce minority representation in Congress. The greatest reductions in minority representation occur in Southern states with large and geographically concentrated Black populations. Finally, these gerrymandered plans further reduce the already low levels of electoral competition in congressional elections by 24 seats on average.
Forthcoming at the Michigan Law Review.
Partisan gerrymandering is a familiar practice. But intra-partisan gerrymandering—one faction within a party designing districts to handicap another faction within the same party—has barely been noticed by courts or scholars. In jurisdictions dominated by a single party, though, intra-partisan gerrymandering is more impactful than its partisan counterpart. In these places, there’s no doubt which party will govern. What’s uncertain is which actors within this party will prevail. Intra-partisan gerrymandering matters precisely because it shapes the identity of the ruling intra-party coalition. In this Article, we first conceptualize intra-partisan gerrymandering. We define the practice, compare it to other forms of abusive redistricting, and explain how it unsettles views of parties as monolithic entities. Next, we provide several examples of intra-partisan gerrymandering. These span the one-party Democratic South, major Democratic cities, and heavily Republican states. We then argue that intra-partisan gerrymandering should be recognized as a distinct legal theory, especially under state constitutions. Its representational harms are analogous to those inflicted by partisan gerrymandering, and it can be regulated through a similar doctrinal framework. Finally, we offer an empirical proof of concept using a recent state senate redistricting as a case study. It’s feasible to identify intra-party factions, measure disparities in their treatment, and assess plans’ fairness with quantitative metrics and computer-generated maps.
Random sampling of redistricting plans has become a standard tool for detecting gerrymandering and evaluating proposed maps. Most MCMC redistricting algorithms operate by merging two adjacent districts and splitting them back into two new districts at each step of the chain. I extend this Merge-Split framework with the Multiple Merge Sequential Split (MMSS) algorithm, which generalizes each step to merge ℓ ≥ 2 adjacent districts and repartition them into ℓ new districts via a sequence of ℓ − 1 uniform spanning tree cuts. MMSS targets the same spanning-forest-weighted distribution used by existing samplers, and I derive the Metropolis-Hastings acceptance ratio. I validate the algorithm by comparing samples against exactly enumerated distributions on small grid maps and compare its scalability with two related algorithms on a realistic redistricting problem. The algorithm supports any value of ℓ ≥ 2, giving users direct control over per-step exploration. MMSS is implemented in the open-source redist R package.
@misc{kenny2026multidistrict,
doi = {10.31235/osf.io/un8sk_v1}},
url = {https://osf.io/preprints/socarxiv/un8sk_v1},
author = {Kenny, Christopher T.},
keywords = {redistricting, gerrymandering, graph partitioning, Markov chain Monte Carlo},
title = {A Multi-District Markov Chain Monte Carlo Sampler for Redistricting},
publisher = {SocArXiv},
year = {2026}
}
Where voters live shapes partisan control of Congress, independently of how district lines are drawn. Political scientists have long argued that Democratic voters concentrate in cities while Republican voters spread more efficiently across suburbs and rural areas, giving Republicans an advantage in redistricting. I show that this geographic advantage has steadily declined. I build a national panel of precinct-level presidential returns from 2008 to 2024, standardized to small, fixed geographic units within each state, and estimate how many House seats Democrats would win under a large sample of nonpartisan congressional redistricting simulations. Under these simulations in a tied national election, the 2008 electorate produces about 203 Democratic seats and the 2024 electorate produces a near-even House. Standard measures of partisan bias in elections similarly show a reduced Republican advantage. However, this decline in geographic partisan advantage does not make the House more competitive. The current distribution of voters by party would produce a fair House absent gerrymandering.
@misc{kenny2026republican,
title = {The Republican geographic advantage in congressional elections has largely disappeared by 2024},
url = {osf.io/preprints/socarxiv/svktd_v1},
publisher = {SocArXiv},
author = {Kenny, Christopher T},
year = {2026},
month = {May}
}
Policymakers in America often hold power over democratic institutions, especially electoral institutions. Their decisions can place principled decision-making at odds with partisan goals, especially in polarized times. How do policymakers balance these competing interests? I argue that policymakers are more likely to side against partisan interests when the law is more explicit. I apply this to partisan gerrymandering in the United States, where map drawers can manipulate district boundaries to favor one party. This provides a hard test, where partisan interests are directly at odds with democratic principles and the stakes of any decision are high. Using new data on the 2020 redistricting cycle combined with redistricting simulations, I find that map drawers typically follow rules that protect partisan fairness. Further, if a partisan gerrymandering case is brought against a redistricting plan, courts are more likely to rule against a plan when there is an explicit law against partisan gerrymandering. When courts intervene, they consistently, but only moderately, decrease the partisan bias of the plan. I then demonstrate that compliance with other, nonpartisan redistricting rules is highest when it is easiest to measure violations. This contributes optimistic evidence that rules effectively bind partisans.
Changes in political geography and electoral district boundaries shape representation in the United States Congress. To disentangle the effects of geography and gerrymandering, we generate a large ensemble of alternative redistricting plans that follow each state's legal criteria. Comparing enacted plans to these simulations reveals partisan bias, while changes in the simulated plans over time identify shifts in political geography. Our analysis shows that geographic polarization has intensified between 2010 and 2020: Republicans improved their standing in rural and rural-suburban areas, while Democrats further gained in urban districts. These shifts offset nationally, reducing the Republican geographic advantage from 14 to 10 seats. Additionally, pro-Democratic gerrymandering in 2020 counteracted earlier Republican efforts, reducing the GOP redistricting advantage by two seats. In total, the pro-Republican bias declined from 16 to 10 seats. Crucially, shifts in political geography and gerrymandering reduced the number of highly competitive districts by over 25%, with geographic polarization driving most of the decline.
@misc{jasny2025gerrymandering,
title = {Gerrymandering and geographic polarization have reduced electoral competition},
author = {Ethan Jasny and Christopher T. Kenny and Cory McCartan and Tyler Simko and Melissa Wu and Michael Y. Zhao and Aneetej Arora and Emma Ebowe and Philip O'Sullivan and Taran Samarth and Kosuke Imai},
year = {2025},
eprint = {2508.15885},
archivePrefix = {arXiv},
primaryClass = {stat.AP},
url = {https://arxiv.org/abs/2508.15885},
}
Bluesky Social is an offshoot of Twitter, designed to provide a decentralized alternative to traditional social media platforms. bskyr is an R package that provides programmatic access to Bluesky Social. The package wraps the official Bluesky Social API, enabling users to retrieve posts, threads, user profiles, social graphs (e.g., follows, mutes, blocks), curated lists, and starter packs. It also supports content creation and interaction so that users can post, reply, like, repost, and follow directly from R. All data is returned in tidy data frames, making it compatible with standard R workflows for analysis and visualization. This allows users of bskyr to not only collect data for observational studies but also to engage with the platform programmatically, such as posting updates or interacting with other users.
Awarded the 2021 Herbert Kaufman Award
Research on inequality overlooks administrative policymaking, where most U.S. law is currently made, under pressure from vast flows of money, lobbying, and political mobilization. Analyzing a new database of over 260,000 comments on agency rules implementing the Dodd-Frank Act, we identify the lobbying activities of over 6,000 organizations. Leveraging measures of organizations' wealth, participation in administrative politics, sophistication, and lobbying success, we provide the first large-scale assessment of wealth-based inequality in agency rulemaking. We find that wealthier organizations are more likely to participate in rulemaking and enjoy more success in shifting the content of federal agency rules. These patterns are not explained by membership differentials. More profit-driven organizations are also more likely to participate and enjoy more success in shifting the content of federal agency rules. Wealthier organizations' ability to marshal legal and technical expertise appears to be a key mechanism by which wealth leads to lobbying success.
No matching items
Publications
2026. Nature Human Behaviour.
Social scientists have developed dozens of measures for assessing partisan bias in redistricting. However, these measures are not easily adapted to other groups, including groups defined by race, class or geography, nor are they applicable to single- or no-party contexts, such as local redistricting. Here we propose a unified framework of harm for evaluating the impacts of a districting plan on individual voters and the groups to which they belong, to overcome these limitations. We consider a voter harmed if their chosen candidate is not elected under the current plan but would be under a different plan. Harm improves on existing measures by both focusing on the choices of individual voters and directly incorporating counterfactual plans. We discuss strategies for estimating harm using redistricting simulations, and demonstrate the utility of our framework through three applications to US redistricting. Overall, harm provides a flexible way to precisely quantify redistricting's individual impacts.
@article{mccartan2026individual,
title = {Individual and differential harm in redistricting},
author = {McCartan, Cory and Kenny, Christopher T.},
journal = {Nature Human Behaviour},
year = {2026},
issn = {2397-3374},
doi = {10.1038/s41562-026-02511-7},
URL = {https://doi.org/10.1038/s41562-026-02511-7},
}
2026. American Political Science Review.
Political actors frequently manipulate redistricting plans to gain electoral advantages, a process commonly known as gerrymandering. To address this problem, several states have implemented institutional reforms including the establishment of map-drawing commissions. It is difficult to assess the impact of such reforms because each state structures bundles of complex rules in different ways. We propose to model redistricting processes as a sequential game. The equilibrium solution to the game summarizes multi-step institutional interactions as a single dimensional score. This score measures the leeway political actors have over the partisan lean of the final plan. Using a differences-in-differences design, we demonstrate that reforms reduce partisan bias and increase competitiveness when they constrain partisan actors. We perform a counterfactual policy analysis to estimate the partisan effects of enacting recent institutional reforms nationwide. We find that instituting redistricting commissions generally reduces the current Republican advantage, but Michigan-style reforms would yield a much greater pro-Democratic effect than types of redistricting commissions adopted in Ohio and New York.
@article{mccartan2026redistricting,
title={Redistricting Reforms Reduce Gerrymandering by Constraining Partisan Actors},
author={McCartan, Cory and Kenny, Christopher T. and Simko, Tyler and Ebowe, Emma and Zhao, Michael Y. and Imai, Kosuke},
journal={American Political Science Review},
year={2026},
pages={1--20},
doi={10.1017/S0003055426101610},
URL={https://doi.org/10.1017/S0003055426101610},
}
2026. Nature Cities.
Residential segregation in US cities threatens economic opportunity and social cohesion. Most segregation estimates rely on aggregate data with arbitrary boundaries such as Census tracts. Sensitivity to boundary choices is well theorized but rarely measured. Here, leveraging redistricting software, we simulate millions of alternative Census tract maps that satisfy Census guidelines, compute segregation for each and estimate probabilistic distributions of common indices. This approach yields new estimates of racial segregation across US cities and measures the aggregation-induced variability hidden in conventional estimates. We find that variability is largely driven by the number of spatial units: simulated segregation in small cities varies widely across plans but converges in larger cities. We find no systematic bias: values calculated using official tract definitions closely track the mean of simulated alternative boundary definitions. While our findings confirm the practical robustness of Census tracts, we provide a general framework for diagnosing and correcting spatial aggregation error in other contexts.
@misc{brown2026city,
author = {Jacob R. Brown and Christopher T. Kenny and Tyler Simko},
title = {City racial segregation statistics are robust to aggregation bias},
journal = {Nature Cities},
year = {2026},
doi = {10.1038/s44284-026-00459-3},
URL = {https://www.nature.com/articles/s44284-026-00459-3},
eprint = {https://www.nature.com/articles/s44284-026-00459-3.epdf},
}
2024. Science Advances.
The United States Census Bureau faces a difficult trade-off between the accuracy of Census statistics and the protection of individual information. We conduct the first independent evaluation of bias and noise induced by the Bureau's two main disclosure avoidance systems: the TopDown algorithm employed for the 2020 Census and the swapping algorithm implemented for the 1990, 2000, and 2010 Censuses. Our evaluation leverages the recent release of the Noisy Measure File (NMF) as well as the availability of two independent runs of the TopDown algorithm applied to the 2010 decennial Census. We find that the NMF contains too much noise to be directly useful alone, especially for Hispanic and multiracial populations. TopDown's post-processing dramatically reduces the NMF noise and produces similarly accurate data to swapping in terms of bias and noise. These patterns hold across census geographies with varying population sizes and racial diversity. While the estimated errors for both TopDown and swapping are generally no larger than other sources of Census error, they can be relatively substantial for geographies with small total populations.
@misc{kenny2024evaluating,
author = {Christopher T. Kenny and Cory McCartan and Shiro Kuriwaki and Tyler Simko and Kosuke Imai},
title = {Evaluating bias and noise induced by the U.S. Census Bureau's privacy protection methods},
journal = {Science Advances},
volume = {10},
number = {18},
pages = {eadl2524},
year = {2024},
doi = {10.1126/sciadv.adl2524},
URL = {https://www.science.org/doi/abs/10.1126/sciadv.adl2524},
eprint = {https://www.science.org/doi/pdf/10.1126/sciadv.adl2524},
}
2024. PNAS.
Current and former Census Bureau officials [Jarmin et al.](https://www.pnas.org/doi/10.1073/pnas.2220558120) argue that differential privacy, which underlies the 2020 Census's Disclosure Avoidance System (DAS), satisfies more desirable theoretical criteria than alternatives. They provide detailed criticisms of many published evaluations of the 2020 DAS, including our work. In this letter, we show that their criticisms are unfounded, grossly mischaracterize our research, and ignore critical issues that merit public discussion.
@article{kenny2024census,
title={Census officials must constructively engage with independent evaluations},
author={Kenny, Christopher T. and McCartan, Cory and Simko, Tyler and Imai, Kosuke},
journal={Proceedings of the National Academy of Sciences},
volume={121},
number={11},
pages={e2321196121},
year={2024},
publisher={National Acad Sciences}
}
2023. PNAS.
Congressional district lines in many U.S. states are drawn by partisan actors, raising concerns about gerrymandering. To isolate the electoral impact of gerrymandering from the effects of other factors including geography and redistricting rules, we compare predicted election outcomes under the enacted plan with those under a large sample of non-partisan, simulated alternative plans for all states. We find that partisan gerrymandering is widespread in the 2020 redistricting cycle, but most of the bias it creates cancels at the national level, giving Republicans two additional seats, on average. In contrast, moderate pro-Republican bias due to geography and redistricting rules remains. Finally, we find that partisan gerrymandering reduces electoral competition and makes the House's partisan composition less responsive to shifts in the national vote.
@article{kenny2023widespread,
author = {Christopher T. Kenny and Cory McCartan and Tyler Simko and Shiro Kuriwaki and Kosuke Imai},
title = {Widespread partisan gerrymandering mostly cancels nationally, but reduces electoral competition},
journal = {Proceedings of the National Academy of Sciences},
volume = {120},
number = {25},
pages = {e2217322120},
year = {2023},
doi = {10.1073/pnas.2217322120},
URL = {https://www.pnas.org/doi/abs/10.1073/pnas.2217322120},
eprint = {https://www.pnas.org/doi/pdf/10.1073/pnas.2217322120},
}
2023. Harvard Data Science Review.
In ["Differential Perspectives: Epistemic Disconnects Surrounding the US Census Bureau's Use of Differential Privacy,"](https://doi.org/10.1162/99608f92.66882f0e) boyd and Sarathy argue that empirical evaluations of the Census Disclosure Avoidance System (DAS), including our published analysis, failed to recognize how the benchmark data against which the 2020 DAS was evaluated is never a ground truth of population counts. In this commentary, we explain why policy evaluation, which was the main goal of our analysis, is still meaningful without access to a perfect ground truth. We also point out that our evaluation leveraged features specific to the decennial Census and redistricting data, such as block-level population invariance under swapping and voter file racial identification, better approximating a comparison with the ground truth. Lastly, we show that accurate statistical predictions of individual race based on the Bayesian Improved Surname Geocoding, while not a violation of differential privacy, substantially increases the disclosure risk of private information the Census Bureau sought to protect. We conclude by arguing that policy makers must confront a key trade-off between data utility and privacy protection, and an epistemic disconnect alone is insufficient to explain disagreements between policy choices.
@article{kenny2023comment,
author = {Kenny, Christopher T. and Kuriwaki, Shiro and McCartan, Cory and Rosenman, Evan T. R. and Simko, Tyler and Imai, Kosuke},
journal = {Harvard Data Science Review},
number = {Special Issue 2},
year = {2023},
month = {jan 31},
note = {https://hdsr.mitpress.mit.edu/pub/6ffzuq19},
publisher = {},
title = {Comment: The {Essential} {Role} of {Policy} {Evaluation} for the 2020 {Census} {DisclosureAvoidance} {System}},
volume = { },
}
2022. Scientific Data.
Covered by The New York Times.
This article introduces the 50stateSimulations, a collection of simulated congressional districting plans and underlying code developed by the Algorithm-Assisted Redistricting Methodology (ALARM) Project. The 50stateSimulations allow for the evaluation of enacted and other congressional redistricting plans in the United States. While the use of redistricting simulation algorithms has become standard in academic research and court cases, any simulation analysis requires non-trivial efforts to combine multiple data sets, identify state-specific redistricting criteria, implement complex simulation algorithms, and summarize and visualize simulation outputs. We have developed a complete workflow that facilitates this entire process of simulation-based redistricting analysis for the congressional districts of all 50 states. The resulting 50stateSimulations include ensembles of simulated 2020 congressional redistricting plans and necessary replication data. We also provide the underlying code, which serves as a template for customized analyses. All data and code are free and publicly available. This article details the design, creation, and validation of the data.
@article{mccartan2022simulated,
title = {Simulated Redistricting Plans for the Analysis and Evaluation of Redistricting in the {{United States}}},
author = {McCartan, Cory and Kenny, Christopher T. and Simko, Tyler and Garcia, George and Wang, Kevin and Wu, Melissa and Kuriwaki, Shiro and Imai, Kosuke},
year = {2022},
month = nov,
journal = {Scientific Data},
volume = {9},
number = {1},
pages = {689},
issn = {2052-4463},
doi = {10.1038/s41597-022-01808-2},
}
2021. Science Advances.
Covered by The Washington Post, Associated Press, NC Policy Watch, and The Harvard Crimson
The US Census Bureau plans to protect the privacy of 2020 Census respondents through its Disclosure Avoidance System (DAS), which attempts to achieve differential privacy guarantees by adding noise to the Census microdata. By applying redistricting simulation and analysis methods to DAS-protected 2010 Census data, we find that the protected data are not of sufficient quality for redistricting purposes. We demonstrate that the injected noise makes it impossible for states to accurately comply with the One Person, One Vote principle. Our analysis finds that the DAS-protected data are biased against certain areas, depending on voter turnout and partisan and racial composition, and that these biases lead to large and unpredictable errors in the analysis of partisan and racial gerrymanders. Finally, we show that the DAS algorithm does not universally protect respondent privacy. Based on the names and addresses of registered voters, we are able to predict their race as accurately using the DAS-protected data as when using the 2010 Census data. Despite this, the DAS-protected data can still inaccurately estimate the number of majority-minority districts. We conclude with recommendations for how the Census Bureau should proceed with privacy protection for the 2020 Census.
@article{kenny2021use,
author = {Christopher T. Kenny and Shiro Kuriwaki and Cory McCartan and Evan T. R. Rosenman and Tyler Simko and Kosuke Imai },
title = {The Use of Differential Privacy for Census Data and its Impact on Redistricting: The Case of the 2020 U.S. Census},
journal = {Science Advances},
volume = {7},
number = {41},
pages = {eabk3283},
year = {2021},
doi = {10.1126/sciadv.abk3283},
URL = {https://www.science.org/doi/abs/10.1126/sciadv.abk3283},
eprint = {https://www.science.org/doi/pdf/10.1126/sciadv.abk3283},
}
2020. Statistics and Public Policy.
As granular data about elections and voters become available, redistricting simulation methods are playing an increasingly important role when legislatures adopt redistricting plans and courts determine their legality. These simulation methods are designed to yield a representative sample of all redistricting plans that satisfy statutory guidelines and requirements such as contiguity, population parity, and compactness. A proposed redistricting plan can be considered gerrymandered if it constitutes an outlier relative to this sample according to partisan fairness metrics. Despite their growing use, an insufficient effort has been made to empirically validate the accuracy of the simulation methods. We apply a recently developed computational method that can efficiently enumerate all possible redistricting plans and yield an independent sample from this population. We show that this algorithm scales to a state with a couple of hundred geographical units. Finally, we empirically examine how existing simulation methods perform on realistic validation datasets.
@article{fifield2020essential,
author = {Benjamin Fifield and Kosuke Imai and Jun Kawahara and Christopher T. Kenny},
title = {The Essential Role of Empirical Validation in Legislative Redistricting Simulation},
journal = {Statistics and Public Policy},
volume = {7},
number = {1},
pages = {52-68},
year = {2020},
publisher = {Taylor & Francis},
doi = {10.1080/2330443X.2020.1791773},
URL = {https://doi.org/10.1080/2330443X.2020.1791773},
eprint = {https://doi.org/10.1080/2330443X.2020.1791773},
}
No matching items
Works-in-progress
Algorithm-Assisted Redistricting Methodology
drafting: earlyBook project
An Individual Causal Framework for Evaluating Electoral Systems
drafting: near completionSocial scientists have developed numerous criteria for evaluating different electoral systems. But these largely depend on aggregate features of the electoral system or its results, not on individual outcomes. Because of this, they are not easily adapted to quantify the impacts of electoral systems on individuals or on groups of individuals defined by race, class, or geography. Moreover, they are ill-suited to capture the effects of changes between electoral systems. To overcome these limitations, we propose a unified causal framework for measuring the effects of electoral reforms on individuals: who benefits, who is harmed, where they live, and what groups they belong to. We define causal measures that zero in on voters whose representational outcomes change as a result of the electoral reform, for better or for worse, and can aggregate these individual gains and losses to quantify differential effects on various groups of voters. The framework and proposed measures improve on existing approaches by both focusing on the choices of individual voters and directly incorporating counterfactual electoral systems, which are always relevant in reform settings. We discuss identification and estimation strategies, and demonstrate the utility of our framework through analyses of voting rights litigation in Alabama, the adoption of ranked-choice voting in Alaska, and redistricting criteria changes in Washington.
Reconsidering the Normal Vote
drafting: near completionNearly all empirical work involving American elections rests on a baseline for how a place “usually” votes. However, the observed vote blends a place's long-term partisanship with short-term forces of a single election, biasing downstream studies. I revive Converse's normal vote, which measures how a group would vote when a given election's short-term forces are in balance, and reformulate it as a property of places rather than of individuals. Applying modern ecological inference to precinct-level presidential returns, I estimate how every precinct in the US would vote in a balanced national environment from 2008 to 2024. I also develop a simpler estimator for settings without a harmonized national panel: shift each election to a tied national vote on the logit scale, then average the adjusted precinct returns. I demonstrate the measure in three applications. First, the normal vote yields an estimate that the incumbency advantage is one to three points. Second, replicating a model of candidate quality, the normal vote improves the model's fit compared to using lagged votes. Third, the normal vote reveals that several 2026 mid-decade gerrymanders improve seat counts for Republicans by targeting recently shifted ground rather than durable partisanship.
The redistverse: A Suite of Packages for Redistricting Analysis
drafting: near completionRedistricting analyses, for both research and litigation, increasingly rely on simulation algorithms to assess the fairness of legislative maps. The redistverse is a family of R packages that provide a comprehensive redistricting analysis pipeline, from data assembly to simulation, analysis, and visualization. The geomander package provides tools for aligning Census data and election returns onto a common precinct geography and for constructing adjacency graphs. The core package, redist, implements Markov chain and sequential Monte Carlo algorithms for generating ensembles of districting plans under realistic legal and geographic constraints. The redistmetrics package provides a suite of metrics for evaluating redistricting plans, including partisan fairness, compactness, and administrative splits. The alarmdata package provides tools for ingesting and reusing data from published research, including data from the Algorithm-Assisted Redistricting Methodology (ALARM) Project Dataverse. The ggredist package provides visualization tools for mapping redistricting plans with ggplot2. Together, these packages provide a robust and reproducible framework for redistricting analysis in R.
Using AI Agents as Research Assistants
drafting: near completionAgentic tooling around large language models (LLMs) allow them to serve as peer programmers. We explore the capabilities of these LLM agents to serve as AI Research Assistants (AIRAs). We explain why researchers should consider AIRAs and provide practical guidance for using them. We then apply AIRAs as research assistants overseeing statistical samplers for redistricting plans from start-to-finish, including making calls about mapping substantive constraints onto mathematical ones. We demonstrate that, like human RAs, AIRAs can take on complex tasks which require making judgements or interpretations. Even with only basic instructions and tutorials, AIRAs can provide outputs that are indistinguishable from that of humans in a fraction of the time and cost. We close with concrete recommendations on using AIRAs as research assistants, including setup and validating outputs. Finally, we discuss how AIRAs are equity-inducing, as they substantially lower the resource barriers for large-scale social science research.
Three Workflows for Verifiable AI-Assisted Research Software
drafting: earlyI discuss and demonstrate three practical workflows for developing statistical software with AI when correctness is the fundamental goal.
Algorithmic Fairness for Redistricting
drafting: earlyApplies broader measures from algorithmic fairness to the harm and benefits research with Cory McCartan to assess if and when standard measures of partisan fairness are applicable in practice.
Measuring Intra-Partisan Gerrymandering
analysis: earlyPolitical science companion piece to our law review article on intra-partisan gerrymandering, developing empirical measures of within-party gerrymandering.
Partisan Effects of House Expansion
analysis: near completionAssesses how expanding the US House would shift partisan seat outcomes, using redistricting simulations to compare equivalently-sized gerrymanders across House sizes.
Four decades of simulated congressional redistricting plans for the United States, 1990-2020
drafting: near completionExtends our simulated redistricting plans for Congress to three more decades, with additional work modernizing old election data and building historical shapefile equivalents for broader use.
No matching items
Public Writing
Case No. 6:25-cv-01980. 2025.
Rebuttal report on simulation analyses for Daytona Beach
Opinion piece in The Hill. 2025.
Case No. 2024 10140 CICI. 2024.
Expert report on census data and map drawing for Daytona Beach
The University of Chicago Center for Effective Government's Democracy Reform Primer Series. 2024.
Alpha Phi Alpha Fraternity, Inc. et al. v. Brad Raffensperger. 2021.
Memo to the Maryland Redistricting Commission. 2021.
No matching items