Causal Inference

AJPS — 2026-09-07 Original Designing multi‐site studies for external validity: Site selection via synthetic purposive sampling

PNAS 123/36 pp. e2530532123-e2530532123 2026-09-03 Original Leveraging generative AI for causal inference with unstructured data Kosuke Imai, Kentaro Nakamura Abstract We introduce GenAI-Powered Inference (GPI), a statistical framework for causal inference using unstructured data, including text and images. GPI leverages open-source pretrained Generative AI (GenAI) models-such as large language models and diffusion models-not only to generate unstructured data at scale but also to extract low-dimensional representations that are guaranteed to capture their underlying structure. Applying machine learning to these representations, GPI enables estimation of causal effects while quantifying estimation uncertainty. Unlike existing approaches to representation learning, GPI does not require fine-tuning of GenAI models, making it computationally efficient and broadly accessible. We illustrate the versatility of the GPI framework through three applications: 1) estimating the effects of Chinese social media censorship while adjusting for textual confounders, 2) isolating the impact of specific image features from that of other correlated features in the same image, and 3) assessing the persuasiveness of political rhetoric. An open-source software package is available for implementing GPI.

AJPS — 2026-07-23 Original Correcting measurement error bias in conjoint survey experiments Katherine Clayton, Yusaku Horiuchi, Aaron R. Kaufman, Gary King, Mayya Komisarchik Abstract Abstract Conjoint survey designs are spreading across the social sciences due to their unusual capacity to estimate many causal effects from a single randomized experiment. Unfortunately, by their ability to mirror complicated real‐world choices, these designs often generate substantial measurement error and thus bias. We replicate both the data collection and analysis from eight prominent conjoint studies, all of which closely reproduce published results, and reveal high levels of measurement error in all. We then discover a common empirical pattern in how measurement error appears in conjoint studies and, with it, introduce an easy‐to‐use statistical method to correct the bias. Along the way, we provide a much simpler and simultaneously more powerful approach to designing, organizing, understanding, and analyzing conjoint data analyses.

JOP 88/3 pp. 1066-1081 2026-07-01 Original Universal Mail Ballot Delivery Boosts Turnout: The Causal Effects of Sending Mail Ballots to All Registered Voters R. Michael Alvarez, Yimeng Li

JPART — 2026-06-15 Original Mapping the temporal evolution of causal effects in public administration and policy research Lefteris Jason Anastasopoulos, Inkyu Kang Abstract Abstract Recent growth in the use of randomized and quasi-experiments in public administration and policy research has advanced the ability to establish cause-and-effect relationships. However, many studies adopt static conceptions of causality, focusing on snapshots or time-averaged effects while overlooking how the effects change over time. This oversight is problematic, as interventions of scholarly interest, such as leadership training or the adoption of new technologies, are likely to produce impacts that unfold in various ways. In this paper, we propose a conceptual framework for understanding the temporal dynamics of causal effects and their implications for research hypotheses and design. We then introduce Bayesian changepoint models (BCMs) as a methodological tool for detecting shifts in the average, variance, or trend of causal effect series, providing a more rigorous yet accessible alternative to visually inspecting graphs. Next, we demonstrate the application of BCMs with an illustrative example derived from simulation data as well as a real-world case examining the effect of body-worn cameras on police officers’ use of force. Finally, we discuss how examining temporal changes in effects can advance theoretical understanding of why and how they occur, as well as inform the design and implementation of policies and strategies in practice.

AJPS — 2026-05-19 Original Post‐instrument bias Julian Schuessler, Adam N. Glynn, Miguel R. Rueda Abstract Abstract When using instrumental variables, researchers often assume that causal effects are only identified conditional on covariates. We show that the role of these covariates is often unclear and that there exists confusion regarding their ability to mitigate violations of the exclusion restriction. We explain when and how existing adjustment strategies may lead to “post‐instrument” bias. We then discuss assumptions that are sufficient to identify various treatment effects when adjustment for post‐instrument variables is required. In general, these assumptions are highly restrictive, albeit they sometimes are testable. We also show that other existing tests are possibly misleading. Then, we introduce a sensitivity analysis that uses information on variables influenced by the instrument to gauge the effect of potential violations of the exclusion restriction. We illustrate it using a published study and summarize our results in easy‐to‐understand guidelines.

APSR — pp. 1-18 2026-04-06 Original Heterogeneous Treatment Effects and Causal Mechanisms JIAWEI FU, TARA SLOUGH Abstract The credibility revolution advances the use of research designs that permit the identification and estimation of causal effects. However, understanding which mechanisms produce measured causal effects remains a challenge. The dominant current approach to the quantitative evaluation of mechanisms relies on the detection of heterogeneous treatment effects (HTEs) with respect to pretreatment covariates. This article develops a framework to understand when the existence of such HTEs can support inferences about the activation of a mechanism. We show first that this design cannot provide evidence of mechanism activation without additional, generally implicit, exclusion assumptions. Further, even when these assumptions are satisfied, the presence of HTEs supports the inference that the mechanism is active but the absence of HTEs is generally uninformative about mechanism activation. We provide novel guidance for interpretation and research design in light of these findings.

PMR — pp. 1-33 2026-03-30 Original Experiments in public administration research: contributions, challenges, and the road ahead Ricardo C. Gomes, Gustavo M. Tavares, Gustavo Mirapalheta Abstract Experiments are essential for testing theory and evaluating policy interventions as they uniquely enable causal inference. This study provides a primer on experimental and quasi-experimental methods and reviews their use in public administration research by analysing 1,143 articles through bibliometric and regression techniques. It focuses on laboratory, field, survey-based, and quasi/natural experiments. The findings reveal rapid growth in experimental research, particularly survey-based experiments, which now dominate the field. Key theoretical clusters, themes, and references are identified, and statistical analyses investigate the link between experiment type and citation counts. The study concludes with recommendations for advancing experimental research in public administration.