Model misspecification and bias for inverse probability weighting and doubly robust estimators

TL;DR

This paper analyzes the bias of semi-parametric causal effect estimators under model misspecification, comparing inverse probability weighting and doubly robust methods, and provides conditions for their relative bias performance.

Contribution

It introduces an analytical approach to study large sample bias of estimators under all models being misspecified and derives conditions for when doubly robust estimators outperform IPW.

Findings

Doubly robust estimators often have smaller bias under moderate outcome model misspecification.

All biases are affected by propensity score model errors, emphasizing careful modeling.

Conditions are derived for when DR estimators outperform simple IPW estimators.

Abstract

In the causal inference literature an estimator belonging to a class of semi-parametric estimators is called robust if it has desirable properties under the assumption that at least one of the working models is correctly specified. In this paper we propose a crude analytical approach to study the large sample bias of semi-parameteric estimators of the average causal effect when all working models are misspecified. We apply our approach to three prototypical estimators, two inverse probability weighting (IPW) estimators, using a misspecified propensity score model, and a doubly robust (DR) estimator, using misspecified models for the outcome regression and the propensity score. To analyze the question of when the use of two misspecified models are better than one we derive necessary and sufficient conditions for when the DR estimator has a smaller bias than a simple IPW estimator and when it has a smaller bias than an IPW estimator with normalized weights. If the misspecificiation of the outcome model is moderate the comparisons of the biases of the IPW and DR estimators suggest that the DR estimator has a smaller bias than the IPW estimators. However, all biases include the PS-model error and we suggest that a researcher is careful when modeling the PS whenever such a model is involved.