7.2 Covariates & Missing pot. outcome perspective
- To estimate the causal effect for any particular unit we need to predict/impute the missing potential outcome
Q: Assumping we have a treatment D (Aspirin) and outcome Y (Pain), what kind of person would you look for to imput Peter (75kg/1.85m/25years)?
We can use pre-treatment variables to make predictions (fill in missings/find comparable units)