Sharp Sensitivity Analysis for Inverse Propensity Weighting via Quantile Balancing

TL;DR

This paper develops a robust sensitivity analysis method for inverse propensity weighting that accounts for unobserved confounding, providing tighter bounds on treatment effects than previous approaches.

Contribution

It introduces a refined sensitivity analysis technique based on quantile balancing, improving the accuracy of treatment effect bounds under unmeasured confounding.

Findings

Provides narrower bounds for treatment effects compared to previous methods

Converges to the narrowest possible interval under given assumptions

Based on new partial identification results for sensitivity models

Abstract

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 paper introduces a robust sensitivity analysis for IPW that estimates the range of treatment effects compatible with a given amount of unobserved confounding. The estimated range converges to the narrowest possible interval (under the given assumptions) that must contain the true treatment effect. Our proposal is a refinement of the influential sensitivity analysis by Zhao, Small, and Bhattacharya (2019), which we show gives bounds that are too wide even asymptotically. This analysis is based on new partial identification results for Tan (2006)'s marginal sensitivity model.