Mediation Analysis Without Sequential Ignorability: Using Baseline Covariates Interacted with Random Assignment as Instrumental Variables

Mediation analysis with principal stratification.

In assessing the mechanism of treatment efficacy in randomized clinical trials, investigators often perform mediation analyses by analyzing if the significant intent-to-treat treatment effect on outcome occurs through or around a third intermediate or mediating variable: indirect and direct effects, respectively. Standard mediation analyses assume sequential ignorability, i.e. conditional on co...

Evaluation of Imputation of Covariates in an Impact Analysis With Regression Adjustment

In an impact analysis using random assignment, researchers often deal with missing values in both the covariates and the outcome variables of regression models. Clearly rigorous methods are needed to impute missing values in the outcome variables to minimize the potential bias in impact assessments. When imputation is applied to covariates of the regression analyses, the effect of imputation is...

Mediation Analysis for Count and Zero-Inflated Count DataWithout Sequential Ignorability and Its Appli- cation in Dental Studies

A Formal Treatment of Sequential Ignorability

Taking a rigorous formal approach, we consider sequential decision problems involving observable variables, unobservable variables, and action variables. We can typically assume the property of extended stability, which allows identification (by means of "[Formula: see text]-computation") of the consequence of a specified treatment strategy if the "unobserved" variables are, in fact, observed-b...

Identification and Estimation of Causal Mediation Effects with Treatment Noncompliance∗

Treatment noncompliance, a common problem in program evaluation, poses a serious challenge to the identification of causal mechanisms via causal mediation analysis. This is because the mediated portion of an intention-to-treat (ITT) effect cannot be nonparametrically identified even when there is no unobserved confounding. In particular, the commonly-used naïve approach of ignoring the actual t...