About

Stephen Ansolabehere is the Frank G. Thompson Professor of Government at Harvard University. He is an expert in public opinion and elections, with extensive publications on elections, mass media, representation, political economy, and public opinion, particularly in relation to energy and the environment. He has authored five books: Cheap and Clean, The Media Game, Going Negative, American Government, and The End of Inequality. Ansolabehere has been recognized as a Carnegie Scholar in 2000, a Hoover National Fellow in 1994, and a Truman Scholar in 1982. He was inducted into the American Academy of Arts and Sciences in 2007. He directed the Caltech/MIT Voting Technology Project from its founding in 2000 through 2004 and serves on the Board of Overseers of the Reuters Institute of Journalism at Oxford University. Additionally, he consults for CBS News Election Decision Desk. At Harvard, he is the director of the Center for American Political Studies and the principal investigator of the Cooperative Congressional Election Study, a collaborative project involving over 60 universities and colleges in the United States.

Research topics

• Political Science
• Computer Science
• Sociology
• Economics
• Artificial Intelligence
• Data Mining
• Business
• Law
• Public economics
• Natural Language Processing
• Advertising
• Psychology
• Public relations
• Law and economics
• Civil engineering
• Epistemology
• Public administration
• Data science
• Engineering
• Social psychology

Selected publications

• City-Defined Neighborhood Boundaries in the United States

  Scientific Data · 2025-06-19 · 2 citations

  articleOpen access1st author

  Neighborhoods are frequently cited as impactful for social, economic, political, and health outcomes. Measuring neighborhoods, however, is challenging, as the definition of a neighborhood may change dramatically across places. Researchers lack widespread but locally-sourced data on neighborhoods, and instead often adopt widely available but arbitrary Census geographies as neighborhood proxies. Others invest in the collection of more precise definitions, but these types of data are hard to collect at scale. We address this tension between scale and precision by collecting, cleaning, and providing to researchers a new dataset of city-defined neighborhoods. Our data includes 206 of the largest cities in the United States, covering more than 77 million people. We combine these data with block-level Census demographic data and provide them along with open-source software to aid researchers in their use.

  PublisherOA PDFDOI

• City-Defined Neighborhood Boundaries in the United States

  2025-03-26

  preprintOpen access1st authorCorresponding

  Neighborhoods are frequently cited as impactful for social, economic, political, and health outcomes. Measuring neighborhoods, however, is challenging, as the definition of a neighborhood may change dramatically across places. Researchers lack widespread but locally-sourced data on neighborhoods, and instead often adopt widely available but arbitrary Census geographies as neighborhood proxies. Others invest in the collection of more precise definitions, but these types of data are hard to collect at scale. We address this tension between scale and precision by collecting, cleaning, and providing to researchers a new dataset of city-defined neighborhoods. Our data includes 206 of the largest cities in the United States, covering more than 77 million people. We combine these data with block-level Census demographic data and provide them along with open-source software to aid researchers in their use.

  PublisherOA PDFDOI

• Collective Representation in Congress

  Perspectives on Politics · 2025-08-27 · 3 citations

  articleOpen access1st authorCorresponding

  The aspiration of representative democracy is that the legislature will make decisions that reflect what the majority of people want. The US Constitution, however, created a Congress with both majoritarian and counter-majoritarian forces. We study public opinion on 103 important issues on the congressional agenda from 2006 to 2022 using the Cooperative Congressional Election Study. Congress made decisions that aligned with what the majority of people wanted on 55% of these issues. Analysis of each issue further reveals the circumstances under which Congress represents the majority and the many ways that representation fails. The likelihood that the House passes a bill is usually a reflection of public support for that policy, but Senate passage depends on how divided the public is on the issue and whether party control of the two chambers of Congress is divided. Legislative institutions make it difficult to pass popular bills but even more difficult to pass unpopular ones. As a result, most representational failures occur because Congress failed to pass a popular bill, rather than because it passed a bill that the public did not want.

  PublisherOA PDFDOI

• Forecasting Turnout

  Harvard Data Science Review · 2024-10-17

  articleOpen access1st authorCorresponding

  We evaluate the predictive power of the leading explanatory models of turnout in the academic literature. We compare the power of using registration, lagged vote, demographics, electoral competition, and early vote data to predict turnout. We specify models to capture each of these approaches to understanding turnout, fit those models to the relevant data from prior elections, and use the estimated parameters from prior years and the relevant observable data from the day of the election in the current year to predict that year’s election. The simplest and most naive model, the Registration Model, out-performed other models in predicting 2016 turnout using 2012 election data and 2020 turnout using 2016 election data. These findings are consistent with classic understandings of which factors most drive turnout, and demonstrate that in modern elections the propensity of registered voters to turnout in presidential elections is fairly stable. Saturated models that combine many of these predictors are common in the academic literature that attempts to explain levels of turnout. We find that such saturated models overfit the data and lead to less accurate predictions than parsimonious models.

  PublisherOA PDFDOI

• Eating Law

  Cambridge University Press eBooks · 2023-06-29

  book-chapter1st authorCorresponding

  In 2014, two titans in the food industry squared off in the Supreme Court of the United States. POM Wonderful, LLC is a well-known beverage producer, largely credited with ushering in America’s love affair with pomegranate juice. POM Wonderful produces a number of pomegranate-based beverages. One such beverage is a “Pomegranate Blueberry” juice that consists primarily of, well, pomegranate and blueberry juice. Not to be left out of the bourgeoning pomegranate juice market, the Coca-Cola Company began manufacturing and selling its own version of a pomegranate blueberry juice drink: Minute Maid Enhanced Pomegranate Blueberry Flavored 100% Juice Blend. It consists of 99.4 percent apple juice.

  PublisherDOI

• The Geography of Racially Polarized Voting: Calibrating Surveys at the District Level

  American Political Science Review · 2023-06-27 · 19 citations

  articleOpen accessCorresponding

  Debates over racial voting, and over policies to combat vote dilution, turn on the extent to which groups’ voting preferences differ and vary across geography. We present the first study of racial voting patterns in every congressional district (CD) in the United States. Using large-sample surveys combined with aggregate demographic and election data, we find that national-level differences across racial groups explain 60% of the variation in district-level voting patterns, whereas geography explains 30%. Black voters consistently choose Democratic candidates across districts, whereas Hispanic and white voters’ preferences vary considerably across geography. Districts with the highest racial polarization are concentrated in the parts of the South and Midwest. Importantly, multiracial coalitions have become the norm: in most CDs, the winning majority requires support from non-white voters. In arriving at these conclusions, we make methodological innovations that improve the precision and accuracy when modeling sparse survey data.

  PublisherOA PDFDOI

• Replication Data for: The Geography of Racially Polarized Voting: Calibrating Surveys at the District Level

  Harvard Dataverse · 2023-01-31 · 1 citations

  datasetOpen access

  <b>Abstract</b>: Debates over racial voting, and over policies to combat vote dilution, turn on the extent to which groups' voting preferences differ and vary across geography. We present the first study of racial voting patterns in every congressional district in the US. Using large-sample surveys combined with aggregate demographic and election data, we find that national-level differences across racial groups explain 60 percent of the variation in district-level voting patterns, while geography explains 30 percent. Black voters consistently choose Democratic candidates across districts, while Hispanic and White voters’ preferences vary considerably across geography. Districts with the highest racial polarization are concentrated in the parts of the South and Midwest. Importantly, multi-racial coalitions have become the norm: in most congressional districts, the winning majority requires support from minority voters. In arriving at these conclusions, we make methodological innovations that improve the precision and accuracy when modeling sparse survey data.

  PublisherDOI

• Cooperative Election Study Common Content, 2022

  Harvard Dataverse · 2023-03-20 · 28 citations

  datasetOpen access

  This is the final release of the 2022 CES Common Content Dataset. The data includes a nationally representative sample of 60,000 American adults. This release includes the data from the survey, a full guide to the data, and the questionnaires. The dataset includes vote validation performed by TargetSmart. Please consult the guide and the study website (https://cces.gov.harvard.edu/frequently-asked-questions) if you have questions about the study.

  PublisherDOI

• Language Models Trained on Media Diets Can Predict Public Opinion

  arXiv (Cornell University) · 2023 · 23 citations

  • Computer Science
  • Computer Science
  • Artificial Intelligence

  Public opinion reflects and shapes societal behavior, but the traditional survey-based tools to measure it are limited. We introduce a novel approach to probe media diet models -- language models adapted to online news, TV broadcast, or radio show content -- that can emulate the opinions of subpopulations that have consumed a set of media. To validate this method, we use as ground truth the opinions expressed in U.S. nationally representative surveys on COVID-19 and consumer confidence. Our studies indicate that this approach is (1) predictive of human judgements found in survey response distributions and robust to phrasing and channels of media exposure, (2) more accurate at modeling people who follow media more closely, and (3) aligned with literature on which types of opinions are affected by media consumption. Probing language models provides a powerful new method for investigating media effects, has practical applications in supplementing polls and forecasting public opinion, and suggests a need for further study of the surprising fidelity with which neural language models can predict human responses.

  PublisherOA PDFDOI

• Cooperative Election Study (CES) Common Content, 2021

  Harvard Dataverse · 2023-01-01

  datasetOpen access

  CES Common Content, 2021 (n = 25,700). In odd years, the CES runs a common content only post-election wave, and multiple modules. This dataset releases the common content.

  PublisherDOI

Recent grants

• CCES, 2012 Panel Study

  NSF · $510k · 2012–2015

• 2018 CSES

  NSF · $845k · 2018–2020

Frequent coauthors

• Jonathan Rodden

  Stanford University

  360 shared

• James M. Snyder

  74 shared

• Brian Schaffner

  Tufts University

  67 shared

• Erik Snowberg

  University of Utah

  33 shared

• Shiro Kuriwaki

  Yale University

  21 shared

• Shanto Iyengar

  Stanford University

  20 shared

• Gary King

  Harvard University Press

  18 shared

• David M. Konisky

  Indiana University Bloomington

  17 shared

Awards & honors

• Carnegie Scholar (2000)
• Hoover National Fellow (1994)
• Truman Scholar (1982)
• inducted into the American Academy of Arts and Sciences (200…