Effective treatment allocation strategies under partial interference

2025-04-09

AI Summary

This paper addresses the problem of optimizing cluster-level treatment assignment under partial interference—where individual potential outcomes depend on others’ treatment statuses—using covariates such as network position and individual attributes. To accommodate heterogeneous interference effects, we propose the first covariate-dependent randomization framework yielding a novel causal estimand that precisely captures individual-level variation in interference strength. We further develop a formal test for interference homogeneity and enable identification of covariates driving global outcomes. Theoretically, we derive the asymptotic distribution of the proposed estimator; empirically, we establish its finite-sample robustness via weighting-based estimation and simulation studies. Applied to a clustered field experiment on insurance uptake in China, our method successfully identifies key individual characteristics moderating the heterogeneous effects of informational interventions, leading to substantial gains in intervention efficiency.

Technology Category

• Reasoning under Uncertainty: Causality
• Machine Learning: Causal Learning
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

Interference occurs when the potential outcomes of a unit depend on the treatment of others. Interference can be highly heterogeneous, where treating certain individuals might have a larger effect on the population's overall outcome. A better understanding of how covariates explain this heterogeneity may lead to more effective interventions. In the presence of clusters of units, we assume that interference occurs within clusters but not across them. We define novel causal estimands under hypothetical, stochastic treatment allocation strategies that fix the marginal treatment probability in a cluster and vary how the treatment probability depends on covariates, such as a unit's network position and characteristics. We illustrate how these causal estimands can shed light on the heterogeneity of interference and on the network and covariate profile of influential individuals. For experimental settings, we develop standardized weighting estimators for our novel estimands and derive their asymptotic distribution. We design an inferential procedure for testing the null hypothesis of interference homogeneity with respect to covariates. We validate the performance of the estimator and inferential procedure through simulations.We then apply the novel estimators to a clustered experiment in China to identify the important characteristics that drive heterogeneity in the effect of providing information sessions on insurance uptake.

Problem

• Understanding how covariates explain heterogeneous interference effects
• Defining causal estimands under stochastic treatment allocation strategies
• Developing estimators to test interference homogeneity and identify influential covariates

Innovation

• Novel causal estimands for treatment allocation strategies
• Standardized weighting estimators for experimental settings
• Inferential procedure for testing interference homogeneity