Policy design in experiments with unknown interference
For researchers and practitioners designing experiments in networked or clustered settings, this work addresses the challenge of unknown spillover effects, offering practical tools for policy evaluation.
This paper develops experimental designs for estimating and testing policies under unknown interference within large clusters, providing a test for policy optimality and a method for estimating welfare-maximizing treatment rules with strong theoretical guarantees and field experiment implementation.
This paper studies experimental designs for estimation and inference on policies with spillover effects. Units are organized into a finite number of large clusters and interact in unknown ways within each cluster. First, we introduce a single-wave experiment that, by varying the randomization across cluster pairs, estimates the marginal effect of a change in treatment probabilities, taking spillover effects into account. Using the marginal effect, we propose a test for policy optimality. Second, we design a multiple-wave experiment to estimate welfare-maximizing treatment rules. We provide strong theoretical guarantees and an implementation in a large-scale field experiment.