Installation | Input data for analyses | Using AIPW to estimate the average treatment effect | One line version (Method chaining from R6class) | A more detailed tutorial | 1. Create an AIPW object | Use SuperLearner libraries | Use sl3 libraries | 2. Fit the AIPW object | 3. Calculate average treatment effects | Estimate the ATE with propensity scores truncation | Check the balance of propensity scores and inverse probability weights by exposure status after truncation | 4. Calculate average treatment effects among the treated/controls | stratified_fit() fits the outcome model by exposure status while fit() does not. Hence, stratified_fit() must be used to compute ATT/ATC (Kennedy et al. 2015) | Parallelization with future.apply | Use tmle/tmle3 fitted object as input | 1. tmle | 2. tmle3

Create an AIPW object | Decorate with Repeated class | More num_reps vs More k-split? | References: