The doubly robust (DR) estimators are widely utilized in observational studies to estimate the average treatment effect due to their consistency when either the outcome model or propensity score is correctly specified. However, in practice, the DR estimator may exhibit significant estimation variability due to misspecification of the outcome model. This paper introduces an alternative DR estimator to enhance efficiency when the outcome model may be misspecified. Specifically, our estimator achieves the smallest asymptotic variance among all DR estimators when the propensity score model is correctly specified, while retaining the double robustness property. To further improve efficiency, we incorporate additional covariates that have predictive power for the outcome variable but do not affect the treatment variable. Simulation studies demonstrate the superior performance of our proposed estimator compared to existing methods. Additionally, we provide two real data analyses to illustrate the proposed method.
Potential outcomes play a fundamental and important role in many causal inference problems. If the potential-outcome means are identifiable, a series of causal effect measures, including the risk difference, the risk ratio, and the treatment benefit rate, among others, can also be identified. However, current identification and estimation methods for these means often implicitly assume that the collected data for analysis are measured precisely. In many fields such as medicine and economics, the collected variables may be subject to measurement errors, such as medical diagnostic results and individual wage data. Misclassification, as a non-classic measurement error, can lead to severely biased estimates in causal inference. In this paper, we leverage a combined sample to study the identifiability of potential-outcome means corresponding to different treatment levers under a plausible misclassification assumption for the outcome, allowing the misclassification probability to depend on not only the true outcome but also the covariates. Furthermore, we propose the multiply-robust and semiparametric efficient estimators for the means, consistent even under partial misspecification of the observed data law, based on the semiparametric theory framework. The simulation studies and real data analysis demonstrate the satisfactory performance of the proposed method.
Proxy variables are commonly used in causal inference when unmeasured confounding exists. While most existing proximal methods assume a unidirectional causal relationship between two primary variables, many social and biological systems exhibit complex feedback mechanisms that imply bidirectional causality. In this paper, using regression-based models, we extend the proximal framework to identify bidirectional causal effects in the presence of unmeasured confounding. We establish the identification of bidirectional causal effects and develop a sensitivity analysis method for violations of the proxy structural conditions. Building on this identification result, we derive bidirectional two-stage least squares estimators that are consistent and asymptotically normal under standard regularity conditions. Simulation studies demonstrate that our approach delivers unbiased causal effect estimates and outperforms some standard methods. The simulation results also confirm the reliability of the sensitivity analysis procedure. Applying our methodology to a state-level panel dataset from 1985 to 2014 in the United States, we examine the bidirectional causal effects between abortion rates and murder rates. The analysis reveals a consistent negative effect of abortion rates on murder rates, while also detecting a potential reciprocal effect from murder rates to abortion rates that conventional unidirectional analyses have not considered.