🤖 AI Summary
This paper addresses the sensitivity of weighted least squares (WLS) estimators in causal inference to unobserved confounding. We propose a general, assumption-lean sensitivity analysis framework grounded in omitted-variable bias theory. Our key innovation is the introduction of the *weighted partial R²*, a measure that quantifies both the direction and magnitude of bias induced by unobserved confounders on weighted causal estimates. From this, we derive a distribution-free sensitivity statistic and an interpretable bound on the strength of unmeasured confounding—expressed as a multiplier relative to observed covariates. The method accommodates arbitrary weighting schemes—including inverse probability weighting (IPW), augmented IPW (AIPW), and stabilized weights—without requiring parametric model specifications or distributional assumptions. An accompanying R package, *weightsense*, enables automated sensitivity reporting and adjusted inference. By providing a transparent, reproducible, and broadly applicable tool for assessing robustness, our framework significantly advances the reliability evaluation of weighted causal estimators.
📝 Abstract
This paper introduces tools for assessing the sensitivity, to unobserved confounding, of a common estimator of the causal effect of a treatment on an outcome that employs weights: the weighted linear regression of the outcome on the treatment and observed covariates. We demonstrate through the omitted variable bias framework that the bias of this estimator is a function of two intuitive sensitivity parameters: (i) the proportion of weighted variance in the treatment that unobserved confounding explains given the covariates and (ii) the proportion of weighted variance in the outcome that unobserved confounding explains given the covariates and the treatment, i.e., two weighted partial $R^2$ values. Following previous work, we define sensitivity statistics that lend themselves well to routine reporting, and derive formal bounds on the strength of the unobserved confounding with (a multiple of) the strength of select dimensions of the covariates, which help the user determine if unobserved confounding that would alter one's conclusions is plausible. We also propose tools for adjusted inference. A key choice we make is to examine only how the (weighted) outcome model is influenced by unobserved confounding, rather than examining how the weights have been biased by omitted confounding. One benefit of this choice is that the resulting tool applies with any weights (e.g., inverse-propensity score, matching, or covariate balancing weights). Another benefit is that we can rely on simple omitted variable bias approaches that, for example, impose no distributional assumptions on the data or unobserved confounding, and can address bias from misspecification in the observed data. We make these tools available in the weightsense package for the R computing language.