Variable Selection for Mediators under a Bayesian Mediation Model

This study proposes a Bayesian variable selection approach to select mediators and quantify their respective posterior probabilities in exploratory mediation analysis. Monte Carlo simulation studies demonstrate that the proposed method has high statistical power selecting mediating effects low Type I error rate excluding null effects. By estimating probability of given effect via distribution, quantifies variable’s influence on continuum scale. is an attractive unique gain neither conventional p-value-based methods nor regularization-based LASSO for possess. We offer four decision rules assist minimize common problem (i.e., elevated type errors) context, as well provide empirical example illustrate method’s application interpretation. end with discussion work directions future work.