Id: arxiv/2603.06941 Url: https://arxiv.org/abs/2603.06941 Tags: math.ST, stat.ME Score: 0 Title: Demonstration Experiments Domain: math.ST Source: arxiv Authors: Guido Imbens, Lorenzo Masoero, Alexander Rakhlin, Thomas S. Richardson, Suhas Vijaykumar Reviews: 0 Subject: math.ST Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-03-06 Page: https://delphi.science/r/arxiv/2603.06941 Adaptive experiments are used extensively in online platforms, healthcare and biotechnology, and the social sciences. Often, the primary goal is not to precisely estimate a treatment effect but to demonstrate that at least one candidate intervention yields a positive effect, for some subpopulation and on some measured outcome. We formalize this objective as testing the global null in a threshold bandit framework, and develop two inference procedures that are valid under general adaptive sampling: one that pools information across promising arms, and one based on time-uniform multiple testing…