Semiparametric Causal Mediation Analysis in Cluster-randomized Experiments

In cluster-randomized experiments, there is emerging interest in exploring the causal mechanism in which a cluster-level treatment affects the outcome through an intermediate variable. Despite an extensive development of causal mediation methods in the past decade, only a few exceptions have been considered in assessing causal mediation in cluster-randomized studies, all of which depend on parametric model-based estimators. In this talk, I will introduce a semiparametric efficiency framework with doubly robust estimators for studying several mediation effect estimands in cluster-randomized experiments, including the natural indirect effect, natural direct effect, and spillover mediation effect. I derive the efficient influence function for each mediation effect, and carefully parameterize each efficient influence function to motivate practical strategies for operationalizing each estimator. I consider both parametric working models and data-adaptive machine learners to estimate the nuisance functions, and obtain semiparametric efficient causal mediation estimators in the latter case. The methods are illustrated via simulations and completed cluster-randomized experiments. The talk is mainly based on the manuscript available at https://arxiv.org/abs/2404.18256.