ohines/plmed: Causal Mediation Effects For Partially Linear Models

This package can be used to estimate causal mediation effects (natural direct and indirect effects) in models where the outcome is modelled by a linear model which include mediator and exposure as covariates. The current package includes G-estimation methods, non-parametric influence curve based methods (TTS) and methods based on Ordinary Least Squares (OLS). Each of these require different combinations of mediator and exposure models, as described in the documentation.

install.packages("remotes")
remotes::install_github("ohines/plmed")