SPER Advanced Methods Workshop, April 2026

Speakers: Jessica G Young, Lan Wen, Soren Harnois-Leblanc

The material included in this repository is destined for participants of the SPER Advanced Methods only. We kindly ask you to not share the code to peers before the paper is published.

• The "SPERworkshop_Gcomp and IPW.R" illustrates the adaptation of g-computation and IPW estimators under a pragmatic intervention g based on the natural value of the exposure.

• The "SPERworkshop_aipw.R" and "SPERworkshop_aipw_ss_pragmatic.R" illustrate the adaptation of the aipw estimator under the pragmatic intervention, with and without the use of machine learning using SuperLearner ensemble models.

• The codes for g-computation, IPW and AIPW will provide values for the estimand. 95% CI need to be estimated with boostrap. We provide the "Boostrap example.R" code for this purpose.

• To run "SPERworkshop_Gcomp and IPW.R" and "SPERworkshop_aipw.R", you will need to download the parametric_data.csv simulated dataset.

• To run the "SPERworkshop_aipw_ss_pragmatic.R", you will need to downnload the ML_data.csv