Multiple Randomization Designs: Estimation and Inference with Interference Article Swipe

PDF

Related Concepts

Estimator Spillover effect Inference Econometrics Computer science Randomized experiment Limit (mathematics) Central limit theorem Sample (material) Value (mathematics) Interference (communication) Mathematics Statistics Economics Artificial intelligence Machine learning Microeconomics Telecommunications Channel (broadcasting) Chromatography Chemistry Mathematical analysis

Lorenzo Masoero , Suhas Vijaykumar , Thomas Richardson , James M. McQueen , Ido M. Rosen , Brian Burdick , Pat Bajari , Guido W. Imbens ·

YOU? · · 2024 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2401.01264 · OA: W4390572814

Classical designs of randomized experiments, going back to Fisher and Neyman in the 1930s still dominate practice even in online experimentation. However, such designs are of limited value for answering standard questions in settings, common in marketplaces, where multiple populations of agents interact strategically, leading to complex patterns of spillover effects. In this paper, we discuss new experimental designs and corresponding estimands to account for and capture these complex spillovers. We derive the finite-sample properties of tractable estimators for main effects, direct effects, and spillovers, and present associated central limit theorems.