Optimizing Cluster-based Randomized Experiments Under Monotonicity

Authors:
Jean Pouget-Abadie Harvard University
David Parkes Harvard University
Vahab Mirrokni Google
Edoardo M. Airoldi Harvard University

Introduction:

This paper studies Cluster-based randomized experiments. The authors introduce a monotonicity condition under which a novel two-stage experimental design allows us to determine which of two cluster-based designs yields the least biased estimator.

Abstract:

Cluster-based randomized experiments are popular designs for mitigating the bias of standard estimators when interference is present and classical causal inference and experimental design assumptions (such as SUTVA or ITR) do not hold. Without an exact knowledge of the interference structure, it can be challenging to understand which partitioning of the experimental units is optimal to minimize the estimation bias. In the paper, we introduce a monotonicity condition under which a novel two-stage experimental design allows us to determine which of two cluster-based designs yields the least biased estimator. We then consider the setting of online advertising auctions and show that reserve price experiments satisfy the monotonicity condition and the proposed framework and methodology apply. We validate our findings on an advertising auction dataset.

You may want to know: