Semiparametric Theory for Causal Mediation Analysis: efficiency bounds, multiple robustness, and sensitivity analysis

By E. T. Tchetgen Tchetgen, Ilya Shpitser. 2012

326 citations recorded by the source corpus.

Whilst estimation of the marginal (total) causal effect of a point exposure on an outcome is arguably the most common objective of experimental and observational studies in the health and social sciences, in recent years, investigators have also become increasingly interested in mediation analysis. Specifically, upon evaluating the total effect of the exposure, investigators routinely wish to make inferences about the direct or indirect pathways of the effect of the exposure not through or through a mediator variable that occurs subsequently to the exposure and prior to the outcome. Although powerful semiparametric methodologies have been developed to analyze observational studies, that produce double robust and highly efficient estimates of the marginal total causal effect, similar methods for mediation analysis are currently lacking. Thus, this paper develops a general semiparametric framework for obtaining inferences about so-called marginal natural direct and indirect causal effects, while appropriately accounting for a large number of pre-exposure confounding factors for the exposure and the mediator variables. Our analytic framework is particularly appealing, because it gives new insights on issues of efficiency and robustness in the context of mediation analysis. In particular, we propose new multiply robust locally efficient estimators of the marginal natural indirect and direct causal effects, and develop a novel double robust sensitivity analysis framework for the assumption of ignorability of the mediator variable.

Subfields: Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Learning Optimal Fair Policies (2018, background): “A clear example of such an estimator for mediation problems appears in (Tchetgen Tchetgen and Shpitser 2012b).”
Fair Inference on Outcomes (2017, methodology): “For the special case of the NDE, one such estimator is the IPW estimator, described earlier and in (Tchetgen and Shpitser 2012b).”
Identification and Sensitivity Analysis for Multiple Causal Mechanisms: Revisiting Evidence from Framing Experiments (2013, background): “Tchetgen Tchetgen and Shpitser (2011) propose a semiparametric sensitivity analysis for the possible existence of unmeasured pre- and posttreatment confounders under Assumption 1 rather than the sensitivity analysis for the possible interaction under Assumption 2. ðtÞ ¼ 3 þ EðMijTi ¼ tÞ þ t ffiffiff”
On Partial Identification of the Pure Direct Effect (2015, background): “While some work has been done to develop sensitivity analyses for unmeasured confounding of the mediator (Tchetgen Tchetgen, 2011; Tchetgen Tchetgen and Shpitser, 2012; Vansteelandt and VanderWeele, 2012), sensitivity analyses for ranges of plausible associations between cross-world counterfactuals…”
Imputation Strategies for the Estimation of Natural Direct and Indirect Effects (2012, methodology): “Relative to our proposal, the approaches of van der Laan and Petersen (2008) and Tchetgen Tchetgen and Shpitser (2011) minimize the risk of incoherent model specification and promise a more guaranteed robustness against model misspecification.”
Mediation Analysis with Multiple Mediators (2014, methodology): “The causal inference literature has provided important advances in mediation analysis in clarifying the assumptions needed for a causal interpretation of direct and indirect effect estimates and in developing approaches to mediation analysis allowing for models with interactions and non-linearities ”
Bias-Reduced Doubly Robust Estimation (2015, methodology): “The content of this introductory chapter is primarily based on the excellent book Semiparametric Theory and Missing Data by Tsiatis A.A. (2006), Springer: New York, which gives a more detailed overview of the semiparametric efficiency theory in missing data problems.”
CAUSAL INFERENCE WITH A GRAPHICAL HIERARCHY OF INTERVENTIONS (2014, methodology): “U eff Ps ( ( a, a′ )) = I(A = a)p(M|a ′,C) p(a|C)p(M|a,C) { Y −E(Y |C,M,a) + I(A = a ′) p(a′|C) { E(Y |C,M,a) − Υa, a′,C + Υa, a′,C − a, a′, where I(·) is the indicator function for an event [31].”
Semiparametric doubly robust targeted double machine learning: a review (2022, methodology): “Under a diﬀerent eﬀect homogeneity assumption, the related but diﬀerent ratio estimand identiﬁes the average treatment eﬀect [Wang and Tchetgen Tchetgen, 2018].”