Kosuke Imai's Research Programs

Bayesian Methods [AI generated]

Bayesian and likelihood-based methods thread through much of my work on political and social measurement. An early contribution developed fast MCMC algorithms for the multinomial probit model and the accompanying MNP R package, enabling scalable Bayesian analysis of discrete choice data. A central application is ideal point estimation: the emIRT algorithm scales Bayesian IRT models to legislative corpora with millions of observations. For network data, Bayesian mixture models underpin a regression framework for dynamic stochastic blockmodels applied to international conflicts and a bipartite network model for legislative cosponsorship. In the context of trade data, a dynamic Bayesian clustering method summarizes billions of product-level observations into interpretable trade profiles. Bayesian inference also plays a central role in my work on sensitive survey questions: models for list experiments, endorsement experiments, randomized response, and auxiliary-information approaches all rely on Bayesian estimation, as do applications comparing these methods in the field and validating them empirically. More recently, Bayesian nonparametrics underlies methods for keyword-assisted topic models, predicting race from voter records, and estimating racial disparities when race is unobserved.

Dasanaike, Noah and Kosuke Imai (2026). ``Using Embedding Models to Improve Probabilistic Race Prediction.''

Lo, Adeline, Santiago Olivella, and Kosuke Imai ``A Statistical Model of Bipartite Networks: Application to Cosponsorship in the United States Senate.'' Political Analysis, Forthcoming.

McCartan, Cory, Robin Fisher, Jacob Goldin, Daniel E. Ho, Kosuke Imai (2025). ``Estimating Racial Disparities When Race is Not Observed.'' Journal of the American Statistical Association, Vol. 120, No. 552, pp. 2140-2153.

Eshima, Shusei, Kosuke Imai, and Tomoya Sasaki (2024). ``Keyword-Assisted Topic Models.'' American Journal of Political Science, Vol. 68, No. 2 (April), pp. 730-750.

Olivella, Santiago, Tyler Pratt, and Kosuke Imai (2022). ``Dynamic Stochastic Blockmodel Regression for Network Data: Application to International Militarized Conflicts.'' Journal of the American Statistical Association, Vol. 117, No. 539, pp. 1068-1081.

Chou, Winston, Kosuke Imai, and Bryn Rosenfeld (2020). ``Sensitive Survey Questions with Auxiliary Information.'' Sociological Methods & Research, Vol. 49, No. 2 (May), pp. 418-454.

Kim, In Song, Steven Liao, and Kosuke Imai (2020). ``Measuring Trade Profile with Granular Product-level Trade Data.'' American Journal of Political Science, Vol. 64, No. 1 (January), pp. 102-117.

Blair, Graeme, Winston Chou, and Kosuke Imai (2019). ``List Experiments with Measurement Error.'' Political Analysis, Vol. 27, No. 4 (October), pp. 455-480.

Hirose, Kentaro, Kosuke Imai, and Jason Lyall (2017). ``Can Civilian Attitudes Predict Insurgent Violence?: Ideology and Insurgent Tactical Choice in Civil War.'' Journal of Peace Research, Vol. 51, No. 1 (January), pp. 47-63. Winner of the Nils Petter Gleditsch Article of the Year Award. Story by Princeton's communication office.

Imai, Kosuke, James Lo, and Jonathan Olmsted (2016). ``Fast Estimation of Ideal Points with Massive Data.'' American Political Science Review, Vol. 110, No. 4 (December), pp. 631-656.

Rosenfeld, Bryn, Kosuke Imai, and Jacob Shapiro (2016). ``An Empirical Validation Study of Popular Survey Methodologies for Sensitive Questions.'' American Journal of Political Science, Vol. 60, No. 3 (July), pp. 783-802.

Imai, Kosuke and Kabir Khanna (2016). ``Improving Ecological Inference by Predicting Individual Ethnicity from Voter Registration Record.'' Political Analysis, Vol. 24, No. 2 (Spring), pp. 263-272.

Blair, Graeme, Kosuke Imai, and Yang-Yang Zhou (2015). ``Design and Analysis of the Randomized Response Technique.'' Journal of the American Statistical Association, Vol. 110, No. 511 (September), pp. 1304-1319.

Lyall, Jason, Yuki Shiraito, and Kosuke Imai (2015). ``Coethnic Bias and Wartime Informing.'' Journal of Politics, Vol. 77, No. 3 (July), p. 833-848.

Imai, Kosuke, Bethany Park, and Kenneth Greene (2015). ``Using the Predicted Responses from List Experiments as Explanatory Variables in Regression Models.'' Political Analysis, Vol. 23, No. 2 (Spring), pp. 180-196. Translated in Portuguese and Reprinted in Revista Debates Vol. 9, No 1.

Blair, Graeme, Kosuke Imai, and Jason Lyall (2014). ``Comparing and Combining List and Endorsement Experiments: Evidence from Afghanistan.'' American Journal of Political Science, Vol. 58, No. 4 (October), pp. 1043-1063.

Lyall, Jason, Graeme Blair, and Kosuke Imai (2013). ``Explaining Support for Combatants during Wartime: A Survey Experiment in Afghanistan.'' American Political Science Review, Vol. 107, No. 4 (November), pp. 679-705. Winner of the Pi Sigma Alpha Award.

Imai, Kosuke, and Dustin Tingley (2012). ``A Statistical Method for Empirical Testing of Competing Theories.'' American Journal of Political Science, Vol. 56, No. 1 (January), pp. 218-236.

Blair, Graeme and Kosuke Imai (2012). ``Statistical Analysis of List Experiments.'' Political Analysis, Vol. 20, No. 1 (Winter), pp. 47-77.

Bullock, Will, Kosuke Imai, and Jacob Shapiro (2011). ``Statistical Analysis of Endorsement Experiments: Measuring Support for Militant Groups in Pakistan.'' Political Analysis, Vol. 19, No. 4 (Autumn), pp. 363-384. (lead article)

Imai, Kosuke (2011). ``Multivariate Regression Analysis for the Item Count Technique.'' Journal of the American Statistical Association, Vol. 106, No. 494 (June), pp. 407-416. (featured article)

Imai, Kosuke, and David A. van Dyk (2005). ``MNP: R Package for Fitting the Multinomial Probit Model.'' Journal of Statistical Software, Vol. 14, No. 3 (May), pp. 1-32.

Imai, Kosuke, and David A. van Dyk (2005). ``A Bayesian Analysis of the Multinomial Probit Model Using Marginal Data Augmentation.'' Journal of Econometrics, Vol. 124, No. 2 (February), pp. 311-334.

Causal Mediation [AI generated]

Understanding why a treatment works—identifying the mechanisms through which causal effects operate—requires methods that go beyond average treatment effects. My research on causal mediation analysis provides a rigorous nonparametric framework for this problem. Foundational papers established identification conditions and sensitivity analysis for average natural direct and indirect effects, and introduced a general approach that unifies and extends the classical Baron–Kenny framework. Experimental designs for identifying causal mechanisms—including crossover and parallel encouragement designs—were developed alongside the theory and read before the Royal Statistical Society. Further work addressed multiple mediators and framing experiments, efficient nonparametric estimation, and sensitivity to contagion effects in placebo-controlled trials. Accessible overviews aimed at applied researchers unpacked the black box of causality and clarified common misunderstandings about mediation and policy. The mediation R package implements these methods and is described in the companion Journal of Statistical Software article and R tutorial.

Imai, Kosuke, and Zhichao Jiang (2020). ``Identification and Sensitivity Analysis of Contagion Effects in Randomized Placebo-Controlled Trials.'' Journal of the Royal Statistical Society, Series A (Statistics in Society), Vol. 183, No. 4 (October), pp. 1637-1657.

Imai, Kosuke, Luke Keele, Dustin Tingley, and Teppei Yamamoto (2014). ``Comment on Pearl: Practical Implications of Theoretical Results for Causal Mediation Analysis.'' Psychological Methods, Vol. 19, No. 4 (December), 482-487.

Tingley, Dustin, Teppei Yamamoto, Luke Keele, and Kosuke Imai (2014). ``mediation: R Package for Causal Mediation Analysis.'' Journal of Statistical Software, Vol. 59, No. 5 (August), pp. 1-38.

Imai, Kosuke and Teppei Yamamoto (2013). ``Identification and Sensitivity Analysis for Multiple Causal Mechanisms: Revisiting Evidence from Framing Experiments.'' Political Analysis, Vol. 21, No. 2 (Spring), pp. 141-171. (lead article)

Imai, Kosuke, Dustin Tingley, and Teppei Yamamoto (2013). ``Experimental Designs for Identifying Causal Mechanisms.'' (with discussions) Journal of the Royal Statistical Society, Series A (Statistics in Society), Vol. 176, No. 1 (January), pp. 5-51. (lead article) Read before the Royal Statistical Society in March, 2012.

Imai, Kosuke, Luke Keele, Dustin Tingley, and Teppei Yamamoto (2011). ``Unpacking the Black Box of Causality: Learning about Causal Mechanisms from Experimental and Observational Studies.'' American Political Science Review, Vol. 105, No. 4 (November), pp. 765-789. Reprinted in Advances in Political Methodology, R. Franzese, Jr. ed., Edward Elger, 2017.

Imai, Kosuke, Luke Keele, and Dustin Tingley (2010). ``A General Approach to Causal Mediation Analysis.'' Psychological Methods, Vol. 15, No. 4 (December), pp. 309-334. (lead article)

Imai, Kosuke, Luke Keele, Dustin Tingley, and Teppei Yamamoto (2010). ``Causal Mediation Analysis Using R.'' ,'' in Advances in Social Science Research Using R, ed. H. D. Vinod, New York: Springer (Lecture Notes in Statistics), pp. 129-154.

Imai, Kosuke, Luke Keele, and Teppei Yamamoto (2010). ``Identification, Inference, and Sensitivity Analysis for Causal Mediation Effects.'' Statistical Science, Vol. 25, No. 1 (February), pp. 51-71.

Chan, K.C.G, K. Imai, S.C.P. Yam, Z. Zhang ``Efficient Nonparametric Estimation of Causal Mediation Effects.''

Ecological Inference [AI generated]

Ecological inference—estimating individual-level relationships from aggregate data—is a fundamental problem whenever individual records are unavailable. My work developed Bayesian and likelihood methods for 2×2 ecological tables using an incomplete-data representation that handles the non-identifiability of the problem gracefully, implemented in the eco R package. An early application examined whether illegal overseas absentee ballots affected the 2000 U.S. presidential election, combining ecological inference with precinct-level data to answer a high-stakes empirical question. A different form of ecological inference arises in the prediction of individual race and ethnicity from administrative records: the wru package uses Bayesian improved surname geocoding (BISG) to predict race from voter registration files, with applications to studies of racial disparity in elections and the criminal justice system.

Dasanaike, Noah and Kosuke Imai (2026). ``Using Embedding Models to Improve Probabilistic Race Prediction.''

Imai, Kosuke, Ying Lu, and Aaron Strauss (2011). ``eco: R Package for Ecological Inference in 2 x 2 Tables.'' Journal of Statistical Software, Vol. 42, No. 5 (Special Volume on Political Methodology), pp. 1-23.

Imai, Kosuke, Ying Lu, and Aaron Strauss (2008). ``Bayesian and Likelihood Inference for 2 x 2 Ecological Tables: An Incomplete Data Approach.'' Political Analysis, Vol. 16, No. 1 (Winter), pp. 41-69.

Imai, Kosuke, and Gary King (2004). ``Did Illegal Overseas Absentee Ballots Decide the 2000 U.S. Presidential Election?.'' Perspectives on Politics, Vol. 2, No. 3 (September), pp.537-549. Our analysis is a part of The New York Times article, ``How Bush Took Florida: Mining the Overseas Absentee Vote'' By David Barstow and Don van Natta Jr. July 15, 2001, Page 1, Column 1.

Experimental Analysis [AI generated]

Randomized experiments are the gold standard for causal inference, but their analysis raises subtle statistical challenges that are often underappreciated. My work addresses several of these. Early contributions examined GOTV field experiments and showed how statistical method choices can reverse substantive conclusions, and developed randomization inference for natural experiments using ballot order in California elections. Paired comparisons in survey experiments and cluster-randomized trials require design-aware analysis that exploits the randomization structure. More recent work has studied priming bias and post-treatment bias in experimental designs, interference and noncompliance in two-stage randomized trials, direct and spillover effect estimation and power analysis, and design-based inference for generalized network experiments with stochastic interventions. Applied experimental work includes an evaluation of economic assistance and combatant support in Afghanistan, a large-scale randomized evaluation of health insurance in India, an evaluation of algorithm-assisted pretrial decision-making, and a study of summer bridge programs for first-generation college students.

Shi, Wenqi, Kosuke Imai, and Yi Zhang (2026). ``Privacy-preserving Meta-analysis through Low-Rank Basis Hunting.''

Sengupta, Souhardya, Kosuke Imai, and Georgia Papadogeorgou (2025). ``Low-rank Covariate Balancing Estimators under Interference.''

Li, Michael Lingzhi and Kosuke Imai (2025). ``Statistical Performance Guarantee for Subgroup Identification with Generic Machine Learning.''

Jiang, Zhichao, Eli Ben-Michael, D. James Greiner, Ryan Halen, Kosuke Imai, and Zhichao Jiang (2025). ``Longitudinal Causal Inference with Selective Eligibility.''

Chattopadhyay, Ambarish, Kosuke Imai, and Jose R. Zubizarreta (2024). ``Design-based inference for generalized network experiments with stochastic interventions.''

Malani, Anup, Phoebe Holtzman, Kosuke Imai, Cynthia Kinnan, Morgen Miller, Shailender Swaminathan, Alessandra Voena, Bartosz Woda, and Gabriella Conti (2021). ``Effect of Health Insurance in India: A Randomized Controlled Trial.''

Blackwell, Matthew, Jacob R. Brown, Sophie Hill, Kosuke Imai, and Teppei Yamamoto (2025). ``Priming bias versus post-treatment bias in experimental designs.'' Political Analysis, Vol. 33, No. 4 (October), pp. 361-377.

Goplerud, Max, Kosuke Imai, Nicole E. Pashley (2025). ``Estimating Heterogeneous Causal Effects of High-Dimensional Treatments: Application to Conjoint Analysis.'' Annals of Applied Statistics, Vol. 19, No. 2 (June), pp. 866-888.

Imai, Kosuke and Michael Lingzhi Li (2025). ``Statistical Inference for Heterogeneous Treatment Effects Discovered by Generic Machine Learning in Randomized Experiments.'' Journal of Business & Economic Statistics, Vol. 43, No. 1, pp. 256-268.

Ham, Dae Woong, Kosuke Imai, and Lucas Janson (2024). ``Using Machine Learning to Test Causal Hypotheses in Conjoint Analysis.'' Political Analysis, Vol. 32, No. 3 (July), pp. 329-344.

Johnson, Rebecca A., Tyler Simko, and Kosuke Imai (2024). ``A Summer Bridge Program for First-Generation Low-Income Students Stretches Academic Ambitions with No Adverse Impacts on GPA.'' Proceedings of the National Academy of Sciences, Vol. 121, No. 50, e2404924121.

Jiang, Zhichao, Kosuke Imai, and Anup Malani (2023). ``Statistical Inference and Power Analysis for Direct and Spillover Effects in Two-Stage Randomized Experiments.'' Biometrics, Vol. 79, No. 3 (September), pp. 2370-2381.

Imai, Kosuke and Michael Lingzhi Li (2023). ``Experimental Evaluation of Individualized Treatment Rules.'' Journal of the American Statistical Association, Vol. 118, No. 541, pp. 242-256.

de la Cuesta, Brandon, Naoki Egami, and Kosuke Imai (2022). ``Improving the External Validity of Conjoint Analysis: The Essential Role of Profile Distribution.'' Political Analysis, Vol. 30, No. 1 (January), pp. 19-45.

Imai, Kosuke, Zhichao Jiang, and Anup Malani (2021). ``Causal Inference with Interference and Noncompliance in Two-Stage Randomized Experiments.'' Journal of the American Statistical Association, Vol. 116, No. 534, pp. 632-644.

Imai, Kosuke, Gary King, and Carlos Velasco Rivera (2020). ``Do Nonpartisan Programmatic Policies Have Partisan Electoral Effects? Evidence from Two Large Scale Randomized Experiments.'' Journal of Politics, Vol. 82, No. 2 (April), pp. 714-730.

Egami, Naoki, and Kosuke Imai (2019). ``Causal Interaction in Factorial Experiments: Application to Conjoint Analysis.'' Journal of the American Statistical Association, Vol. 114, No. 526 (June), pp. 529-540.

Imai, Kosuke and Marc Ratkovic (2013). ``Estimating Treatment Effect Heterogeneity in Randomized Program Evaluation.'' Annals of Applied Statistics, Vol. 7, No. 1 (March), pp. 443-470. Winner of the Tom Ten Have Memorial Award. Reprinted in Advances in Political Methodology, R. Franzese, Jr. ed., Edward Elger, 2017.

Imai, Kosuke, and Aaron Strauss (2011). ``Estimation of Heterogeneous Treatment Effects from Randomized Experiments, with Application to the Optimal Planning of the Get-out-the-vote Campaign.'' Political Analysis, Vol. 19, No. 1 (Winter), pp. 1-19. (lead article) Winner of Political Analysis Editors' Choice Award.

Imai, Kosuke (2005). ``Do Get-Out-The-Vote Calls Reduce Turnout? The Importance of Statistical Methods for Field Experiments.'' American Political Science Review, Vol. 99, No. 2 (May), pp. 283-300.

Experimental Design [AI generated]

Careful experimental design is as important as the analysis that follows. My research on experimental design spans several settings. For factorial and conjoint experiments—widely used in political science and marketing—I developed theory for causal interaction effects and the role of profile distribution in external validity, as well as methods for testing causal hypotheses via machine learning and estimating heterogeneous effects of high-dimensional treatments. For experiments aimed at identifying causal mechanisms, a framework of experimental designs—crossover, parallel encouragement, and related variants—provides the tools to estimate mediation pathways. Work on cluster-randomized experiments showed that pair matching is essential for efficiency and that ignoring the matched structure leads to conservative inference; variance identification under the matched-pair design provides the supporting theory. The regression discontinuity design has also received attention, with work clarifying common misunderstandings in its application to close elections. Further work on priming bias and post-treatment bias clarifies important design pitfalls in survey experiments.

Ham, Dae Woong, Kosuke Imai, and Lucas Janson (2024). ``Using Machine Learning to Test Causal Hypotheses in Conjoint Analysis.'' Political Analysis, Vol. 32, No. 3 (July), pp. 329-344.

Imai, Kosuke, and Zhichao Jiang (2018). ``A Sensitivity Analysis for Missing Outcomes Due to Truncation-by-Death under the Matched-Pairs Design.'' Statistics in Medicine, Vol. 37, No. 20 (September), pp. 2907-2922.

Imai, Kosuke, and Teppei Yamamoto (2010). ``Causal Inference with Differential Measurement Error: Nonparametric Identification and Sensitivity Analysis.'' American Journal of Political Science, Vol. 54, No. 2 (April), pp. 543-560.

Imai, Kosuke (2009). ``Statistical Analysis of Randomized Experiments with Nonignorable Missing Binary Outcomes: An Application to a Voting Experiment.'' Journal of the Royal Statistical Society, Series C (Applied Statistics), Vol. 58, No. 1 (February), pp. 83-104.

Ho, Daniel E., and Kosuke Imai (2008). ``Estimating Causal Effects of Ballot Order from a Randomized Natural Experiment: California Alphabet Lottery, 1978-2002.'' Public Opinion Quarterly, Vol. 72, No. 2 (Summer), pp. 216-240.

Imai, Kosuke (2008). ``Sharp Bounds on the Causal Effects in Randomized Experiments with ``Truncation-by-Death.'' Statistics & Probability Letters, Vol. 78, No. 2 (February), pp. 144-149.

Horiuchi, Yusaku, Kosuke Imai, and Naoko Taniguchi (2007). ``Designing and Analyzing Randomized Experiments: Application to a Japanese Election Survey Experiment.'' American Journal of Political Science, Vol. 51, No. 3 (July), pp. 669-687.

Ho, Daniel E., and Kosuke Imai (2006). ``Randomization Inference with Natural Experiments: An Analysis of Ballot Effects in the 2003 California Recall Election.'' Journal of the American Statistical Association, Vol. 101, No. 475 (September), pp. 888-900.

Heterogeneous Treatment Effects [AI generated]

Estimating how treatment effects vary across individuals—and using those estimates to make better decisions—is a central theme of my research. Foundational work developed nonparametric methods for heterogeneous effects in randomized experiments with application to optimal GOTV campaigns, and introduced support vector machine-based estimators for treatment effect heterogeneity. The generic machine learning (GML) framework provides valid inference for heterogeneous effects discovered by any black-box machine learning algorithm in randomized experiments, including subgroup identification with statistical guarantees. For conjoint data, methods for causal interaction in factorial experiments, high-dimensional treatments, and machine learning-based hypothesis testing address effect heterogeneity in multi-attribute designs. Extensions include regret-based evaluation of individualized treatment rules, on-policy bandit evaluation, and heterogeneous effects of airstrikes on insurgent violence in spatio-temporal settings. Work on Neyman-meets-ML evaluation methods connects classical experimental inference with modern machine learning.

Zhou, Lingxiao, and Kosuke Imai, Jason Lyall, and Georgia Papadogeorgou (2025). ``Estimating Heterogeneous Treatment Effects for Spatio-Temporal Causal Inference: How Economic Assistance Moderates the Effects of Airstrikes on Insurgent Violence.''

Li, Michael Lingzhi and Kosuke Imai (2025). ``Statistical Performance Guarantee for Subgroup Identification with Generic Machine Learning.''

Jia, Zeyang, Kosuke Imai, and Michael Lingzhi Li (2025). ``Cramming Contextual Bandits for On-policy Statistical Evaluation.''

Zhang, Yi and Kosuke Imai (2025). ``Individualized Policy Evaluation and Learning under Clustered Network Interference.''

Zhang, Yi, Melody Huang, and Kosuke Imai (2024). ``Minimax Regret Estimation for Generalizing Heterogeneous Treatment Effects with Multisite Data.''

Ham, Dae Woong, Kosuke Imai, and Lucas Janson (2024). ``Using Machine Learning to Test Causal Hypotheses in Conjoint Analysis.'' Political Analysis, Vol. 32, No. 3 (July), pp. 329-344.

Li, Michael Lingzhi and Kosuke Imai (2024). ``Neyman Meets Causal Machine Learning: Experimental Evaluation of Individualized Treatment Rules.'' Journal of Causal Inference, Vol 12, No. 1, pp. 1-20. Special Issue on Neyman (1923) and its influences on causal inference

Imai, Kosuke and Michael Lingzhi Li (2023). ``Experimental Evaluation of Individualized Treatment Rules.'' Journal of the American Statistical Association, Vol. 118, No. 541, pp. 242-256.

Machine Learning & AI [AI generated]

Machine learning methods offer powerful tools for causal inference and social science measurement, but require careful adaptation to ensure valid statistical guarantees. My research integrates ML into causal inference pipelines in several ways. The generic machine learning (GML) framework uses any supervised learner to discover heterogeneous treatment effects, then applies sample-splitting and asymptotic theory to obtain valid confidence intervals—extended to subgroup identification, individualized treatment rule evaluation, regret minimization, and on-policy bandit evaluation. In conjoint analysis, ML methods underpin hypothesis tests for causal effects and estimation of high-dimensional treatment effects. For text and image data, keyword-assisted topic models combine supervised keywords with unsupervised LDA, automated coding of campaign advertisement videos demonstrates scalable measurement from video, and AI-assisted summarization of the full 1952–2012 CMAG video archive shows large-scale application. Most recently, causal representation learning with generative AI uses large language models to construct causal representations of textual treatments, and Gen-AI powered inference develops a general framework for using generative AI to assist statistical inference tasks that would otherwise require extensive human annotation.

Dasanaike, Noah and Kosuke Imai (2026). ``Using Embedding Models to Improve Probabilistic Race Prediction.''

Shi, Wenqi, Kosuke Imai, and Yi Zhang (2026). ``Privacy-preserving Meta-analysis through Low-Rank Basis Hunting.''

Imai, Kosuke and Kentaro Nakamura (2025). ``Gen-AI Powered Inference.''

Imai, Kosuke and Kentaro Nakamura (2025). ``Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments.''

Li, Michael Lingzhi and Kosuke Imai (2025). ``Statistical Performance Guarantee for Subgroup Identification with Generic Machine Learning.''

Jia, Zeyang, Kosuke Imai, and Michael Lingzhi Li (2025). ``Cramming Contextual Bandits for On-policy Statistical Evaluation.''

Zhang, Yi and Kosuke Imai (2025). ``Individualized Policy Evaluation and Learning under Clustered Network Interference.''

Zhang, Yi, Melody Huang, and Kosuke Imai (2024). ``Minimax Regret Estimation for Generalizing Heterogeneous Treatment Effects with Multisite Data.''

Breuer, Adam, Bryce J. Dietrich, Michael H. Crespin, Matthew Butler, J.A. Pyrse, Kosuke Imai (2025). ``Using AI to Summarize US Presidential Campaign TV Advertisement Videos, 1952-2012.'' Scientific Data, Vol. 12, No. 1552.

Ham, Dae Woong, Kosuke Imai, and Lucas Janson (2024). ``Using Machine Learning to Test Causal Hypotheses in Conjoint Analysis.'' Political Analysis, Vol. 32, No. 3 (July), pp. 329-344.

Eshima, Shusei, Kosuke Imai, and Tomoya Sasaki (2024). ``Keyword-Assisted Topic Models.'' American Journal of Political Science, Vol. 68, No. 2 (April), pp. 730-750.

Tarr, Alexander, June Hwang, and Kosuke Imai (2023). ``Automated Coding of Political Campaign Advertisement Videos: An Empirical Validation Study.'' Political Analysis, Vol. 31, No. 4 (October), pp. 554-574.

Imai, Kosuke and Michael Lingzhi Li (2023). ``Experimental Evaluation of Individualized Treatment Rules.'' Journal of the American Statistical Association, Vol. 118, No. 541, pp. 242-256.

Ning, Yang, Sida Peng, and Kosuke Imai (2020). ``Robust Estimation of Causal Effects via High-Dimensional Covariate Balancing Propensity Score.'' Biometrika, Vol. 107, No. 3 (September), pp. 533-554.

Svyatkovskiy, Alexey, Kosuke Imai, Mary Kroeger, and Yuki Shiraito (2016). ``Large-scale text processing pipeline with Apache Spark.'' IEEE International Conference on Big Data, Washington, DC, pp. 3928-3935.

Matching & Weighting [AI generated]

A central challenge in causal inference from observational data is adjusting for confounding without imposing strong parametric assumptions. My research on matching and weighting addresses this challenge across a range of settings. Early work introduced matching as nonparametric preprocessing for parametric causal inference, showing that preprocessing data with methods such as those implemented in the MatchIt software package can substantially reduce model dependence and improve the robustness of causal estimates, and clarified common misunderstandings about the relationship between experimental and observational approaches. A parallel line of work developed the Covariate Balancing Propensity Score (CBPS), which estimates propensity scores by directly optimizing covariate balance—yielding more robust inverse probability weights for marginal structural models, continuous treatments, high-dimensional settings, and with optimal theoretical guarantees. A related contribution generalized the propensity score to arbitrary treatment regimes, with extensions to non-binary treatments. For panel and time-series cross-sectional data, I showed that unit fixed effects regression models implicitly perform a form of matching, developed analogous results for two-way fixed effects models, and built the PanelMatch framework to make these connections explicit. Additional work addressed matched-pairs designs in randomized experiments—including cluster-randomized trials and sensitivity analysis for missing outcomes—and applied these ideas to evaluate large-scale social programs. Recent work extends these methods to spatio-temporal settings where the treatment is a point process in space and time, with extensions to heterogeneous treatment effects and the geocausal package for spillover and carryover effects. Further ongoing work develops low-rank covariate balancing estimators under network interference and distributionally robust causal estimators that remain valid under departures from the assumed propensity score model.

Sengupta, Souhardya, Kosuke Imai, and Georgia Papadogeorgou (2025). ``Low-rank Covariate Balancing Estimators under Interference.''

Mukaigawara, Mitsuru, and Kosuke Imai, Jason Lyall, and Georgia Papadogeorgou (2025). ``Spatiotemporal causal inference with arbitrary spillover and carryover effects.''

Bertsimas, Dimitris, Kosuke Imai, and Michael Lingzhi Li (2023). ``Distributionally Robust Causal Inference with Observational Data.''

Tarr, Alexander and Kosuke Imai (2025). ``Estimating Average Treatment Effects with Support Vector Machines.'' Statistics in Medicine, Vol. 44, No. 5, e70006.

Imai, Kosuke, In Song Kim, and Erik Wang (2023). ``Matching Methods for Causal Inference with Time-Series Cross-Sectional Data.'' American Journal of Political Science, Vol. 67, No. 3 (July), pp. 587-605.

Fan, Jianqing, Kosuke Imai, Inbeom Lee, Han Liu, Yang Ning, and Xiaolin Yang (2023). ``Optimal Covariate Balancing Conditions in Propensity Score Estimation.'' Journal of Business & Economic Statistics, Vol. 41, No. 1, pp. 97-110.

Imai, Kosuke, and Yang Ning (2023). ``Imai, Kosuke, and Yang Ning. (2023). ``Covariate Balancing Propensity Score.'' Handbook of Matching and Weighting Adjust.''

Papadogeorgou, Georgia, Kosuke Imai, Jason Lyall, and Fan Li (2022). ``Causal Inference with Spatio-temporal Data: Estimating the Effects of Airstrikes on Insurgent Violence in Iraq.'' Journal of the Royal Statistical Society, Series B (Statistical Methodology), Vol. 84, No. 5 (November), pp. 1969-1999.

Imai, Kosuke and In Song Kim (2021). ``On the Use of Two-way Fixed Effects Regression Models for Causal Inference with Panel Data.'' Political Analysis, Vol. 29, No. 3 (July), pp. 405-415.

Imai, Kosuke and James Lo (2021). ``Robustness of Empirical Evidence for the Democratic Peace: A Nonparametric Sensitivity Analysis.'' International Organization, Vol. 75, No. 3 (Summer), pp. 901-919.

Zhao, Shandong, David A. van Dyk, and Kosuke Imai (2020). ``Propensity-Score Based Methods for Causal Inference in Observational Studies with Non-Binary Treatments.'' Statistical Methods in Medical Research, Vol. 29, No. 3 (March), pp. 709-727.

Imai, Kosuke and In Song Kim (2019). ``When Should We Use Unit Fixed Effects Regression Models for Causal Inference with Longitudinal Data?.'' American Journal of Political Science, Vol. 63, No. 2 (April), pp. 467-490.

Fong, Christian, Chad Hazlett, and Kosuke Imai (2018). ``Covariate Balancing Propensity Score for a Continuous Treatment: Application to the Efficacy of Political Advertisements.'' Annals of Applied Statistics, Vol. 12, No. 1, pp. 156-177.

Imai, Kosuke and Marc Ratkovic (2015). ``Robust Estimation of Inverse Probability Weights for Marginal Structural Models.'' Journal of the American Statistical Association, Vol. 110, No. 511 (September), pp. 1013-1023. (lead article)

Imai, Kosuke and Marc Ratkovic (2014). ``Covariate Balancing Propensity Score.'' Journal of the Royal Statistical Society, Series B (Statistical Methodology), Vol. 76, No. 1 (January), pp. 243-263.

Ho, Daniel E., Kosuke Imai, Gary King, and Elizabeth Stuart (2011). ``MatchIt: Nonparametric Preprocessing for Parametric Causal Inference.'' Journal of Statistical Software, Vol. 42, No. 8 (Special Volume on Political Methodology), pp. 1-28.

King, Gary, Emmanuela Gakidou, Kosuke Imai, Jason Lakin, Ryan T. Moore, Clayton Nall, Nirmala Ravishankar, Manett Vargas, Martha María Téllez-Rojo, Juan Eugenio Hernández Ávila, Mauricio Hernández Ávila, and Héctor Hernández Llamas (2009). ``Public Policy for the Poor? A Randomised Assessment of the Mexican Universal Health Insurance Programme.'' (with a comment) The Lancet, Vol. 373, No. 9673 (April), pp. 1447-1454.

Imai, Kosuke, Gary King, and Clayton Nall (2009). ``Rejoinder: Matched Pairs and the Future of Cluster-Randomized Experiments.'' Statistical Science, Vol. 24, No. 1 (February), pp. 65-72.

Imai, Kosuke, Gary King, and Clayton Nall (2009). ``The Essential Role of Pair Matching in Cluster-Randomized Experiments, with Application to the Mexican Universal Health Insurance Evaluation.'' (with discussions and rejoinder) Statistical Science, Vol. 24, No. 1 (February), pp. 29-53.

Imai, Kosuke (2008). ``Variance Identification and Efficiency Analysis in Randomized Experiments under the Matched-Pair Design.'' Statistics in Medicine, Vol. 27, No. 24 (October), pp. 4857-4873.

Imai, Kosuke, Gary King, and Elizabeth A. Stuart (2008). ``Misunderstandings among Experimentalists and Observationalists about Causal Inference.'' Journal of the Royal Statistical Society, Series A (Statistics in Society), Vol. 171, No. 2 (April), pp. 481-502. Reprinted in Field Experiments and their Critics, D. Teele ed. (2014), New Haven: Yale University Press.

Ho, Daniel E., Kosuke Imai, Gary King, and Elizabeth A. Stuart (2007). ``Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference.'' Political Analysis, Vol. 15, No.3 (Summer), pp. 199-236. (lead article) Winner of Miller Prize.

Imai, Kosuke, and David A. van Dyk (2004). ``Causal Inference With General Treatment Regimes: Generalizing the Propensity Score.'' Journal of the American Statistical Association, Vol. 99, No. 467 (September), pp. 854-866.

Missing Data [AI generated]

Missing data are pervasive in experimental and observational research, and ignoring them can severely bias causal inferences. My work addresses missing data in several experimental contexts. In randomized experiments with truncation by death—where the outcome is undefined for units who do not survive—sharp bounds on causal effects provide the tightest possible nonparametric conclusions. When binary outcomes are missing non-ignorably, a likelihood-based method using instrumental variables yields consistent estimates under plausible assumptions. Measurement error in the treatment variable is addressed by nonparametric identification and sensitivity analysis that bounds the bias from differential misclassification. In matched-pairs designs, a sensitivity analysis for missing outcomes assesses robustness of conclusions to the assumption that missingness is independent of potential outcomes. The design and analysis of randomized experiments framework handles item nonresponse in survey experiments through principled imputation.

Imai, Kosuke (2008). ``Sharp Bounds on the Causal Effects in Randomized Experiments with ``Truncation-by-Death.'' Statistics & Probability Letters, Vol. 78, No. 2 (February), pp. 144-149.

Network & Interference [AI generated]

Standard causal inference assumes no interference between units—that one unit's treatment does not affect another's outcome. My research relaxes this assumption in several important settings. In two-stage randomized experiments, where clusters are first randomized to treatment saturation and then individuals within clusters are randomized, work has developed identification and estimation of direct and spillover effects under noncompliance, power analysis for direct and spillover effects, and sensitivity analysis for contagion effects in placebo-controlled trials. For network experiments more broadly, design-based inference for generalized network experiments with stochastic interventions provides a unified framework that accommodates arbitrary interference structures. Recent work develops low-rank covariate balancing estimators that remain valid under network interference, and distributionally robust causal inference that is robust to misspecification of the interference model. Applied work has estimated spillover effects of health insurance on non-enrolled household members in a large-scale randomized trial in India.

Sengupta, Souhardya, Kosuke Imai, and Georgia Papadogeorgou (2025). ``Low-rank Covariate Balancing Estimators under Interference.''

Chattopadhyay, Ambarish, Kosuke Imai, and Jose R. Zubizarreta (2024). ``Design-based inference for generalized network experiments with stochastic interventions.''

Bertsimas, Dimitris, Kosuke Imai, and Michael Lingzhi Li (2023). ``Distributionally Robust Causal Inference with Observational Data.''

Lo, Adeline, Santiago Olivella, and Kosuke Imai ``A Statistical Model of Bipartite Networks: Application to Cosponsorship in the United States Senate.'' Political Analysis, Forthcoming.

Policy Learning & Evaluation [AI generated]

A natural goal of estimating heterogeneous treatment effects is to use them to improve policy decisions—selecting who receives treatment to maximize welfare. My research on policy learning develops the statistical foundations for this problem. Early work derived optimal GOTV campaign strategies from field experiments using nonparametric methods. Subsequent work formalized experimental evaluation of individualized treatment rules, developed Neyman-meets-ML evaluation methods, and studied regret minimization and on-policy evaluation via contextual bandits. Statistical guarantees for subgroup identification with generic machine learning and inference for heterogeneous effects in randomized experiments underpin these policy applications. Work on individualized policy evaluation under clustered network interference extends the framework to settings where spillovers complicate welfare calculations. Applications span algorithmic decision-making in pretrial public safety assessment, fairness in algorithmic decisions, safety of AI-assisted decisions, asymmetry of power in AI deployment, regression discontinuity-based policy evaluation, and safe policy learning via Bayesian optimization.

Koch, Benedikt and Kosuke Imai (2025). ``Statistical Decision Theory with Counterfactual Loss.''

Li, Michael Lingzhi and Kosuke Imai (2025). ``Statistical Performance Guarantee for Subgroup Identification with Generic Machine Learning.''

Jia, Zeyang, Kosuke Imai, and Michael Lingzhi Li (2025). ``Cramming Contextual Bandits for On-policy Statistical Evaluation.''

Zhang, Yi and Kosuke Imai (2025). ``Individualized Policy Evaluation and Learning under Clustered Network Interference.''

Zhang, Yi, Eli Ben-Michael, and Kosuke Imai (2024). ``Safe Policy Learning under Regression Discontinuity Designs.''

Jia, Zeyang, Eli Ben-Michael, and Kosuke Imai ``Bayesian Safe Policy Learning with Chance Constrained Optimization: Application to Military Security Assessment during the Vietnam War.'' Journal of the Royal Statistical Society, Series A (Statistics in Society), Forthcoming.

Ben-Michael, Eli, D. James Greiner, Kosuke Imai, and Zhichao Jiang (2025). ``Safe Policy Learning through Extrapolation: Application to Pre-trial Risk Assessment.'' Journal of the American Statistical Association, Vol. 120, No. 551, pp. 1386-1399.

Ben-Michael, Eli, D. James Greiner, Melody Huang, Kosuke Imai, Zhichao Jiang, Sooahn Shin (2025). ``Does AI help humans make better decisions? A statistical evaluation framework for experimental and observational studies.'' Proceedings of the National Academy of Sciences, Vol. 122, No. 38, e2505106122.

Ben-Michael, Eli, Kosuke Imai, and Zhichao Jiang (2024). ``Policy Learning with Asymmetric Counterfactual Utilities.'' Journal of the American Statistical Association, Vol. 119, No. 548, pp. 3045-3058.

Imai, Kosuke and Zhichao Jiang (2023). ``Principal Fairness for Human and Algorithmic Decision-Making.'' Statistical Science, Vol. 38, No. 2 (July), pp317-328.

Imai, Kosuke and Michael Lingzhi Li (2023). ``Experimental Evaluation of Individualized Treatment Rules.'' Journal of the American Statistical Association, Vol. 118, No. 541, pp. 242-256.

Imai, Kosuke, Zhichao Jiang, D. James Greiner, Ryan Halen, and Sooahn Shin (2023). ``Experimental Evaluation of Algorithm-Assisted Human Decision-Making: Application to Pretrial Public Safety Assessment.'' (with discussion) Journal of the Royal Statistical Society, Series A (Statistics in Society), Vol. 186, No. 2 (April), pp. 167-189. Read before the Royal Statistical Society.

Racial Disparity [AI generated]

Measuring and correcting for racial disparities in administrative data requires methods that account for the fact that race is often unobserved or imputed. My work has built a suite of tools for this problem. The wru package implements Bayesian improved surname geocoding (BISG) to predict individual race from voter registration files; subsequent work addressed known biases in BISG by incorporating first and middle name information and using a fully Bayesian estimation approach. A companion dataset of race and ethnicity for first, middle, and last names was released to support this research. The birdie package takes a different approach, estimating racial disparities directly from aggregate data when individual race is unobserved—providing consistent estimates under weaker assumptions than imputation-based methods. Together these tools have been applied to audit racial disparities in policing, lending, and electoral administration.

Dasanaike, Noah and Kosuke Imai (2026). ``Using Embedding Models to Improve Probabilistic Race Prediction.''

Rosenman, Evan T.R., Santiago Olivella, and Kosuke Imai (2023). ``Race and ethnicity data for first, middle, and last names.'' Scientific Data, Vol. 10, No. 299, pp. 1-11.

Imai, Kosuke, Santiago Olivella, and Evan T.R. Rosenman (2022). ``Addressing Census data problems in race imputation via fully Bayesian Improved Surname Geocoding and name supplements.'' Science Advances, Vol. 8, No. 49 (December), pp. 1-10.

Record Linkage [AI generated]

Linking records across administrative datasets—identifying which records in two files refer to the same individual—is a prerequisite for many empirical studies in social science and medicine. My work developed a probabilistic model for large-scale administrative record linkage that scales to tens of millions of records and is implemented in the fastLink software package. A key insight is that linkage uncertainty should be propagated into downstream analyses rather than ignored. An application links public opinion surveys to voter files to validate self-reported voter turnout, one of the most widely studied outcomes in political science—finding that survey respondents systematically overreport their turnout relative to official records.

Dasanaike, Noah and Kosuke Imai (2026). ``Using Embedding Models to Improve Probabilistic Race Prediction.''

Enamorado, Ted, Benjamin Fifield, and Kosuke Imai (2019). ``Using a Probabilistic Model to Assist Merging of Large-scale Administrative Records.'' American Political Science Review, Vol. 113, No. 2 (May), pp. 353-371.

Enamorado, Ted, and Kosuke Imai (2019). ``Validating Self-reported Turnout by Linking Public Opinion Surveys with Administrative Records.'' Public Opinion Quarterly, Vol. 83, No. 4 (Winter), pp. 723-748.

Redistricting [AI generated]

Legislative redistricting has profound consequences for electoral competition and representation, yet the analysis of redistricting plans has historically been hampered by the lack of rigorous computational and statistical methods. My research has contributed to both. The redist package implements Markov chain Monte Carlo methods for automated redistricting simulation, and a Sequential Monte Carlo algorithm samples compact and balanced plans more efficiently, with a generalized SMC extension for broader applications. The 50stateSimulations dataset provides simulated redistricting plans for all U.S. states, enabling a nationwide study of partisan gerrymandering showing that partisan bias and electoral competition reduction are widespread but largely cancel nationally. Work on redistricting reforms shows that independent commissions and other institutional changes constrain partisan actors and reduce gerrymandering. Extensions address Japanese redistricting, geographic polarization and electoral competition, and the intersection of redistricting with Census data privacy in a study of the 2020 Census Disclosure Avoidance System and its downstream impact.

O'Sullivan, Philip, Kosuke Imai, and Cory McCartan (2026). ``Generalized Sequential Monte Carlo Sampling for Redistricting Simulation.''

Miyazaki, Sho, Kento Yamada, and Kosuke Imai (2025). ``Estimating the Partisan Bias of Japanese Legislative Redistricting Plans Using a Simulation Algorithm.''

Jasny, Ethan, Christopher T. Kenny, Cory McCartan, Tyler Simko, Melissa Wu, Michael Y. Zhao, Aneetej Arora, Emma Ebowe, Philip O'Sullivan, Taran Samarth, and Kosuke Imai (2025). ``Gerrymandering and geographic polarization have reduced electoral competition.''

McCartan, Cory, Christopher Kenny, Tyler Simko, Emma Ebowe, Michael Zhao, and Kosuke Imai ``Redistricting Reforms Reduce Gerrymandering by Constraining Partisan Actors.'' American Political Science Review, Forthcoming.

McCartan, Cory and Kosuke Imai (2023). ``Sequential Monte Carlo for Sampling Balanced and Compact Redistricting Plans.'' Annals of Applied Statistics, Vol. 17, No. 4 (December), pp. 3300-3323.

Kenny, Christopher T., Cory McCartan, Tyler Simko, Shiro Kuriwaki, and Kosuke Imai (2023). ``Widespread Partisan Gerrymandering Mostly Cancels Nationally, but Reduces Electoral Competition.'' Proceedings of the National Academy of Sciences, Vol. 120, No. 25, e2217322120.

Kenny, Christopher T., Shiro Kuriwaki, Cory McCartan, Evan Rosenman, Tyler Simko, and Kosuke Imai (2023). ``Comment: The Essential Role of Policy Evaluation for the 2020 Census Disclosure Avoidance System.'' Harvard Data Science Review, Special Issue 2: Dierential Privacy for the 2020 U.S. Census (January).

McCartan, Cory, Christopher T. Kenny, Tyler Simko, George Garcia III, Kevin Wang, Melissa Wu, Shiro Kuriwaki, and Kosuke Imai (2022). ``Simulated redistricting plans for the analysis and evaluation of redistricting in the United States.'' Scientific Data, Vol. 9, No. 689, pp. 1-10.

Kenny, Christopher T., Shiro Kuriwaki, Cory McCartan, Evan T.R. Rosenman, Tyler Simko, and Kosuke Imai (2021). ``The Use of Differential Privacy for Census Data and its Impact on Redistricting: The Case of the 2020 U.S. Census.'' Science Advances, Vol. 7, No. 7 (October), pp. 1-17.

Fifield, Benjamin, Kosuke Imai, Jun Kawahara, and Christopher T. Kenny (2020). ``The Essential Role of Empirical Validation in Legislative Redistricting Simulation.'' Statistics and Public Policy, Vol. 7, No. 1, pp 52-68.

Fifield, Benjamin, Michael Higgins, Kosuke Imai, and Alexander Tarr (2020). ``Automated Redistricting Simulation Using Markov Chain Monte Carlo.'' Journal of Computational and Graphical Statistics, Vol. 29, No. 4, pp. 715-728.

Sensitive Questions and Privacy [AI generated]

Asking sensitive questions—about stigmatized behaviors, illegal activities, or controversial opinions—poses a fundamental challenge: respondents may give socially desirable rather than truthful answers. My research has developed and analyzed several indirect questioning techniques that provide statistical privacy guarantees while enabling inference about population proportions. The list experiment (item count technique) hides individual responses within a count, and my work provides multivariate regression analysis, methods for using predicted responses as covariates, and measurement error corrections. The randomized response technique has been redesigned and analyzed from a modern design-based perspective. Endorsement experiments measure support for sensitive groups by embedding them in policy questions, and comparisons across methods in Afghanistan provide empirical guidance for practitioners. Auxiliary information from non-sensitive questions can sharpen inference, and empirical validation studies compare these methods head-to-head. A separate strand of work addresses privacy in Census data: the 2020 Census Disclosure Avoidance System, based on differential privacy, raised serious concerns about its impact on redistricting and research, leading to published letters and responses to Census Bureau rejoinders.

Kenny, Christopher, Cory McCartan, Tyler Simko, and Kosuke Imai (2024). ``Census officials must constructively engage with independent evaluations.'' Proceedings of the National Academy of Sciences (Letter), Vol. 121, No. 11, e2321196121.

Kenny, Christopher, Cory McCartan, Shiro Kuriwaki, Tyler Simko, and Kosuke Imai (2024). ``Evaluating Bias and Noise Induced by the U.S. Census Bureau's Privacy Protection Methods.'' Science Advances, Vol 10, No. 18 (May), pp. 1-13.

McCartan, Cory, Tyler Simko, and Kosuke Imai (2024). ``Rejoinder: We Can Improve the Usability of the Census Noisy Measurements File.'' Harvard Data Science Review, Vol. 6, No. 2 (Spring).

McCartan, Cory, Tyler Simko, and Kosuke Imai (2023). ``Making Differential Privacy Work for Census Data Users.'' Harvard Data Science Review, Vol. 5, No. 4 (Fall).

McCartan, Cory, Tyler Simko, and Kosuke Imai (2023). ``Researchers need better access to US Census data.'' Science, Vol. 380, No. 6648 pp. 902-903.

Chou, Winston, Kosuke Imai, and Bryn Rosenfeld (2020). ``Sensitive Survey Questions with Auxiliary Information.'' Sociological Methods & Research, Vol. 49, No. 2 (May), pp. 418-454.

Blair, Graeme, Winston Chou, and Kosuke Imai (2019). ``List Experiments with Measurement Error.'' Political Analysis, Vol. 27, No. 4 (October), pp. 455-480.

Lyall, Jason, Yuki Shiraito, and Kosuke Imai (2015). ``Coethnic Bias and Wartime Informing.'' Journal of Politics, Vol. 77, No. 3 (July), p. 833-848.

Blair, Graeme and Kosuke Imai (2012). ``Statistical Analysis of List Experiments.'' Political Analysis, Vol. 20, No. 1 (Winter), pp. 47-77.

Imai, Kosuke (2011). ``Multivariate Regression Analysis for the Item Count Technique.'' Journal of the American Statistical Association, Vol. 106, No. 494 (June), pp. 407-416. (featured article)

Spatial Data [AI generated]

Many important social and political phenomena are inherently spatial. My work incorporates spatial structure into causal inference and measurement in several ways. Measuring and modeling neighborhoods uses spatial data to construct granular measures of racial composition and segregation, applied to studying neighborhood effects on political behavior. In conflict research, causal inference with spatio-temporal point process treatments estimates the effects of airstrikes on insurgent violence in Iraq, treating the treatment itself as a spatial point process—implemented in the geocausal package, which also handles arbitrary spillover and carryover effects. Heterogeneous treatment effects in spatio-temporal settings reveal how economic assistance moderates the effects of military operations. Redistricting research relies heavily on spatial data: simulated redistricting plans, sequential Monte Carlo sampling, and geographic polarization analysis all process and analyze fine-grained spatial information about Census blocks and electoral precincts.

O'Sullivan, Philip, Kosuke Imai, and Cory McCartan (2026). ``Generalized Sequential Monte Carlo Sampling for Redistricting Simulation.''

Miyazaki, Sho, Kento Yamada, and Kosuke Imai (2025). ``Estimating the Partisan Bias of Japanese Legislative Redistricting Plans Using a Simulation Algorithm.''

Mukaigawara, Mitsuru, and Kosuke Imai, Jason Lyall, and Georgia Papadogeorgou (2025). ``Spatiotemporal causal inference with arbitrary spillover and carryover effects.''

McCartan, Cory, Jacob Brown, and Kosuke Imai (2024). ``Measuring and Modeling Neighborhoods.'' American Political Science Review, Vol. 118, No. 4 (November), pp. 1966-1985.

McCartan, Cory and Kosuke Imai (2023). ``Sequential Monte Carlo for Sampling Balanced and Compact Redistricting Plans.'' Annals of Applied Statistics, Vol. 17, No. 4 (December), pp. 3300-3323.

Texts Images & Video [AI generated]

Rich unstructured data—text, images, and video—are increasingly available in social science research, but extracting reliable measurements from them requires methods that combine machine learning with principled statistical inference. The keyATM package combines user-provided seed keywords with unsupervised LDA to produce topic models that are both data-driven and interpretable, with Bayesian uncertainty quantification. In the domain of video data, automated coding of political campaign advertisement videos validates machine-coded measures against human coders, and AI-assisted summarization of the full 1952–2012 CMAG video archive demonstrates large-scale application. Two recent papers push the frontier of causal inference with text: causal representation learning with generative AI uses large language models to construct causal representations of textual treatments, enabling credible causal inference about the effects of text-based interventions; and Gen-AI powered inference develops a general framework for using generative AI to assist statistical inference tasks that would otherwise require extensive human annotation.

Imai, Kosuke and Kentaro Nakamura (2025). ``Gen-AI Powered Inference.''

Imai, Kosuke and Kentaro Nakamura (2025). ``Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments.''

Eshima, Shusei, Kosuke Imai, and Tomoya Sasaki (2024). ``Keyword-Assisted Topic Models.'' American Journal of Political Science, Vol. 68, No. 2 (April), pp. 730-750.