Sharp Sensitivity Analysis for Inverse Propensity Weighting via Quantile Balancing

Inverse propensity weighting (IPW) is a popular method for estimating treatment effects from observational data. However, its correctness relies on the untestable (and frequently implausible) assumption that all confounders have been measured. This article introduces robust sensitivity analysis IPW estimates range of compatible with given amount unobserved confounding. The estimated converges to narrowest possible interval (under assumptions) must contain true effect. Our proposal refinement influential by Zhao, Small, and Bhattacharya, which we show gives bounds are too wide even asymptotically. based new partial identification results Tan’s marginal model. Supplementary materials this available online.