Causal Inference

This paper asks whether voters’ affect toward groups made salient by housing policy explains opposition to new development and reforms. Drawing on symbolic politics theory, the authors test whether dislike of developers, investors, government approvers, or incoming residents helps account for anti-development preferences—especially for policies that are not tied to a specific local project. The contribution is experimental and causal: the authors report evidence from survey experiments and short-form video manipulations that link affect toward these actors to policy judgments.

This article develops a clear framework for defining, identifying, and estimating the causal effect of an observed treatment on a latent outcome, labeled the latent treatment effect (LTE). The focus is on latent political concepts such as values, beliefs, and attitudes, and on how experimental or observational treatments can be linked causally to those unobserved traits.

How do we know the extent of racial bias in policing? New research suggests administrative records often mask it. Police officers, when choosing whom to investigate, may show racial discrimination that isn't captured by standard data analysis methods. This gap means past estimates could be seriously off.

This paper investigates whether the arrival of railroads affected the state’s ability to implement national school policy in nineteenth-century Sweden by making school inspectors more able to monitor remote schools. Using fine-grained local education data, the authors compare communities reachable by rail to those that were not and assess differences in school permanence and curricular content. The paper argues for a causal interpretation for at least some outcomes, and presents more tentative evidence on curricular change.

This study asks whether renewable terms compromise judicial independence by creating incentives for judges to curry favor with actors who control their reappointment. Using a 2010 reform at the European Court of Human Rights (ECtHR) that made judges’ terms nonrenewable with immediate effect, the author estimates how removing reappointment opportunities changed judges’ voting behavior on cases involving their nominating states. The paper’s contribution is a focused causal test of the reappointment mechanism versus selection explanations for pro-state voting.

Prespecifying the network of ties is a major obstacle for applied spatial analysis. When networks are unknown or misspecified, researchers face bias from omitted spatial inputs or from errors in the spatial weights matrix.

This paper examines how HRO activities influence human rights conditions. It explores two models: boomerang (where third-party pressure indirectly drives improvements) and spiral (where such pressure can worsen conditions). Using a causal mediation model, it investigates whether the effects of HRO campaigns stem from direct targeting or through induced costs by powerful actors like other states or international organizations via sanctions or military interventions. The findings reveal that while HROs have an overall positive effect on human rights, third-party pressure partially offsets this improvement.

Dropping subjects after random assignment because they fail a manipulation check is common across the social sciences, usually intended to restrict estimates to those thought to understand the prompt. This practice can, however, introduce serious bias and obscures what the experiment actually reveals unless handled differently.

Roll rates measure how often legislators vote with their party's majority. Counterfactual Roll Rate Simulation

Problem:** Civil war studies debate the right fatality threshold but rarely check how results change across different levels. This could miss crucial differences between minor and major conflicts.

Measuring the causal effect of state behavior on outcomes faces two central obstacles: behavior is endogenous, and unobserved confounders are pervasive. Commonly used matching methods are poorly suited when confounders are unobserved. Heckman-style multiple-equation models address unobserved confounding but depend on rigid functional-form assumptions that can introduce substantial bias in estimates of average treatment effects.

This article evaluates Google Consumer Surveys (GCS) as a low-cost option for rigorous social science. The platform’s relative strengths and weaknesses make it most useful for researchers who identify causality through randomization to treatment groups rather than through selection on observables. The real cost advantage of GCS over alternatives is largely confined to short surveys with only a few questions.

Political scientists often convert treatments with multiple intensities (for example, years of schooling) into coarser measures (for example, a binary indicator for completing high school). This common choice can substantially upwardly bias instrumental variable (IV) estimates by subtly violating the exclusion restriction, producing misleading causal inferences.

New methods are presented to estimate causal effects retrospectively from microdata by harnessing a machine learning ensemble. The approach targets a clearly defined "retrospective intervention effect" founded on hypothetical population interventions, so the causal comparisons are specified rather than ambiguous.