Generalizing causal effects with noncompliance: Application to deep canvassing experiments

2025-05-30

AI Summary

This paper addresses the problem of extrapolating causal effects from experimental samples to target populations under noncompliance, focusing on identifying and estimating the target-population complier average causal effect (T-CACE). Conventional instrumental variable (IV) methods fail due to distributional differences between experimental and target populations and rely on the untestable principal ignorability assumption. To overcome these limitations, we propose novel identification conditions based on population-level exclusion restrictions, construct an inverse-probability-weighted estimator incorporating auxiliary compliance information, and develop a sensitivity analysis framework for unmeasured confounding. We establish theoretical guarantees: consistency and asymptotic normality of the estimator. Simulation studies demonstrate high accuracy and robustness across scenarios. Empirical application to a field experiment reveals that in-depth lobbying significantly reduces xenophobic attitudes, and the estimated T-CACE generalizes credibly to broader populations.

Technology Category

• Machine Learning: Causal Learning
• Reasoning under Uncertainty: Causality
• Application Domains: Humanities & Computational Social Science

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Economics, Online Markets and Human Computation: Fairness, privacy, and diversity in economic environments

Abstract

Standard approaches in generalizability often focus on generalizing the intent-to-treat (ITT). However, in practice, a more policy-relevant quantity is the generalized impact of an intervention across compliers. While instrumental variable (IV) methods are commonly used to estimate the complier average causal effect (CACE) within samples, standard approaches cannot be applied to a target population with a different distribution from the study sample. This paper makes several key contributions. First, we introduce a new set of identifying assumptions in the form of a population-level exclusion restriction that allows for identification of the target complier average causal effect (T-CACE) in both randomized experiments and observational studies. This allows researchers to identify the T-CACE without relying on standard principal ignorability assumptions. Second, we propose a class of inverse-weighted estimators for the T-CACE and derive their asymptotic properties. We provide extensions for settings in which researchers have access to auxiliary compliance information across the target population. Finally, we introduce a sensitivity analysis for researchers to evaluate the robustness of the estimators in the presence of unmeasured confounding. We illustrate our proposed method through extensive simulations and a study evaluating the impact of deep canvassing on reducing exclusionary attitudes.

Problem

• Generalizing causal effects for compliers across populations
• Identifying target complier average causal effect without principal ignorability
• Proposing estimators and sensitivity analysis for unmeasured confounding

Innovation

• Introduces population-level exclusion restriction assumptions
• Proposes inverse-weighted estimators for T-CACE
• Develops sensitivity analysis for unmeasured confounding