Estimation of the Average Treatment Effect among Responders
Usage
RATE(
response,
post.treatment,
treatment,
data,
M = 5,
pr.treatment,
treatment.level,
preprocess = NULL,
efficient = TRUE,
...
)Arguments
- response
(formula or learner) Response model. A formula (e.g.,
Y ~ D*A) is wrapped in learner_glm with a Gaussian family.- post.treatment
(formula or learner) Post treatment marker model. A formula (e.g.,
D ~ W) is wrapped in learner_glm with a binomial family.- treatment
Treatment formula (e.g, A ~ 1)
- data
data.frame
- M
Number of folds in cross-fitting (M=1 is no cross-fitting)
- pr.treatment
(optional) Randomization probability of treatment.
- treatment.level
Treatment level in binary treatment (default 1)
- preprocess
(optional) Data preprocessing function
- efficient
If TRUE, the estimate will be efficient. If FALSE, the estimate will be a simple plug-in estimate.
- ...
Additional arguments to lower level functions