Estimate Average Treatment Effect with Bootstrapped CI's
boot_ATE(model, treat, R = 250, block = "", df)
| model | GLM regression object |
|---|---|
| treat | Character string for variable to calculate ATE with |
| R | Integer for number of bootstrap replications |
| block | Character string for "block" to sample by (cluster robust SE's). |
| df | Dataframe object of original sample |
coefficient from glm object
coefficient from bootstraps
ATE (p1- p0)
RR (p1/p0)
matrix of bootstrap replications
boot_ATE will take a glm regression object and compute predicted outcomes with treatment = 1 and treatment 0; It will then bootstrap the model and compute effects. There is an option for block bootstrapping to compute cluster robust intervals. Bootstrapped confidence intervals are percentile.