Causal Identification under Interference: The Role of Treatment Assignment Independence

2026-04-24

AI Summary

This study addresses the violation of the no-interference assumption in causal inference when interference among units is present, which typically induces bias in effect estimation. The authors clarify that under arbitrary interference, standard identification formulas actually target the average direct effect and establish, for the first time, that this effect remains identifiable even without knowledge of the interference structure or exposure mapping, provided treatment assignment is independent of potential outcomes. Building on this result, they develop a sensitivity analysis framework to assess how violations of this independence assumption affect causal conclusions. The proposed approach applies broadly to mainstream causal strategies—including instrumental variables, regression discontinuity, and difference-in-differences—thereby delineating the validity boundaries of conventional methods under interference and offering a practical tool for quantifying robustness.

Technology Category

• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning
• Knowledge Representation and Reasoning: Action, Change, and Causality

Application Category

• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs
• Social Networks and Social Media: Fairness and bias in social network and social media analysis
• Security and Privacy: Large-scale security measurements

Abstract

Empirical researchers routinely invoke the no-interference or \textit{individualistic treatment response} (ITR) assumption to identify causal effects in observational studies, despite concerns that interference across units may arise in many economic settings. This paper studies the causal content of standard ITR-based identification formulas when arbitrary interference is present. We show that, under restrictions on dependence between treatment assignments across units, conventional ITR-based identification formulas -- including those underlying selection-on-observables, instrumental variables, regression discontinuity designs, and difference-in-differences -- identify well-defined causal objects: types of \textit{average direct effects} (ADEs). These results do not require knowledge of the interference structure or specification of exposure mappings. We also propose a sensitivity analysis framework that quantifies the robustness of statistical inference to violations of treatment-assignment independence under arbitrary interference.

Problem

• causal identification
• interference
• treatment assignment independence
• average direct effects
• observational studies

Innovation

• causal identification
• interference
• treatment assignment independence
• average direct effects
• sensitivity analysis