B0964
Title: Causal mediation with instrumental variables
Authors: Kara Rudolph - Columbia University (United States) [presenting]
Ivan Diaz - NYU Langone Health (United States)
Nicholas Williams - Columbia University (United States)
Abstract: Mediation analysis is a strategy for understanding the mechanisms by which interventions affect later outcomes. However, unobserved confounding concerns may be compounded in mediation analyses, as there may be unobserved exposure-outcome, exposure-mediator, and mediator-outcome confounders. Instrumental variables (IVs) are a popular identification strategy in the presence of unobserved confounding. However, in contrast to the rich literature on the use of IV methods to identify and estimate the total effect of a non-randomized exposure, there has been almost no research into using IV as an identification strategy to identify mediational indirect effects. In response, novel estimands are defined and nonparametrically identified, complier interventional direct and indirect effects, with a single IV for the exposure, and when two, possibly related, IVs are available, one for the exposure and another for the mediator. Nonparametric, robust, efficient estimators are proposed for these effects, as well as for related complier natural direct and indirect effects, and those are applied to a housing voucher experiment.
