Introduction to Double Robust Estimation for Causal Inference
-
Please use the "Causal_Setup_Instructions.pdf" to install or update R on the laptop that you will bring to the workshop. This should take 15 minutes of your time at most. The instructions include loading in R packages and reading in the data for the workshop: “CausalWorkshop.csv”. Once you have downloaded/updated R, and read in the data, you are ready for the workshop!
-
If you are interested, feel free to take a look at the exercise: “Causal_Rimplementation.pdf". The exercise is self-contained, but please do not get overwhelmed by the material covered here. We will work through this together, using slides that are less technical, and participants will have many opportunities to ask questions.
Workshop Goals: (a) Introduce the Roadmap for causal inference with an applied example. (b) In R, implement the simple substitution estimator (a.k.a. parametric G-computation) and targeted maximum likelihood estimation (TMLE) with the ensemble algorithm Super Learner for estimation of the effect of a binary exposure. (c) Show the ltmle and drtmle packages in R. (d)Time-permitting: Explore the performance of the estimators with simulations.