About

About this site

Preface to current version

The current tutorial version should show a basic start to finish example of the “Clone-Censor-Weight” approach. It is performed on a simulated dataset with no particular consideration for different types of confounding, or specialized cases etc. The tutorial will be updated when possible based on feedback, advances in science or as solutions to particular problems become known.

The synthetic data was manually created for this project.

Some planned guides:

  1. Aggregation of continuous times into discrete intervals
  2. Assessment of covariate balance in weighted sample
  3. Covariate selection

Versions

Version 0.0.1 02-19-2025

Project set-up

Version 0.1 05-12-2025

  • Synthetic data set-up for observed effect by treatment due to confounding; no effect after appropriate adjustment
  • Estimator file complete; additional edits for clarity and alternative approaches
  • Inference file TBD
  • Advanced file discusses some bootstrap and computation approaches
  • Appendix TBD

Version 0.2 05-13-2025

  • Added Parallel computation step to Advanced Topics page
  • Setup for the inference page (WIP)

Version 0.3 05-21-2025

  • Revamped synthetic data for time-varying confounding
  • Added cox models to estimator page
  • switched to color-blind friendly color-scheme

Version 0.5 05-22-2025

  • Tutorial appears to work as expected, some edits to text and code to be more concise