This is our last module on the design of RCTs. In this module, we’ll learn how to use machine learning to choose a set of baseline covariates to include when we estimate treatment effects, and we’ll also learn how the choice of covariates influences the required sample size.
Blog Post: Using PDS Lasso to Select Controls in Field Experiments
Using Post-Double Selection Lasso in Field Experiments
Optional: An Introduction to Statistical Learning: Chapter 6
A handout version of the lecture slides is available here.
The Stata version of the empirical exercise is available here.