Reevaluating Causal Estimation Methods with Data from a Product Release (opens in new tab)

arXiv:2601.11845v2 Announce Type: replace-cross Abstract: Recent developments in causal machine learning methods have made it easier to estimate flexible relationships between confounders, treatments and outcomes, making unconfoundedness assumptions in causal analysis more palatable. How successful are these approaches in recovering ground truth baselines? In this paper we analyze a new data sample including an experimental rollout of a new feature at a large technology company and a simultan...