Covariate Adjustment with a Binary Outcome: Tutorial Materials

This is meant to provide a worked example of applying standardization (G-computation) to a binary outcome in R.

How to Use This Repo:

1. Make sure you have the latest version of R and R Studio installed.

2. Make sure that your packages are up to date (see update.packages()).

3. If using Windows, install RTools: Toolchains. Mac and Linux should already have compilers available for building packages from source.

4. Open R Studio: Select File > New Project > Version Control > Git:

- In the `Repository URL:` field, paste `https://github.com/CovariateAdjustment/BinaryOutcomeTutorial`
- In the `Project directory name:` field, paste `BinaryOutcomeTutorial`
- Choose a directory for cloning the repo.

5. R Studio will open the project once it is cloned: Run 0_install_required_packages.r: This will install packages from R and CRAN.

After these steps are complete, open Covariate_Adjustment_Handout_Binary: This R Markdown report provides a step-by-step example of materials in the slides.

---

Content in Repository

Data

1. `Simulated_MISTIE_III_v1.2.csv” - A spreadsheet of the data used in the example
2. sim_MIII_MRS.Rdata - Materials used in handouts and slides
3. sim_MIII_MRS_fixed.Rdata - Cached results used in handouts and slides

Helper Functions

1. boot_p_value.R contains boot_p_value() which can take a boot object produced by boot::boot() and calculate a p-value by finding the smallest confidence level $\alpha$ at which the null hypothesis is rejected (i.e. the CI no longer contains the null value of the parameter).

Covariate Adjustment

• Covariate_Adjustment_Slides_Binary contains the slides used in short courses
• Covariate_Adjustment_Handout_Binary shows a reproducible report that walks through all of the code examples in the slides.

Additional Code

In order to reduce compiling/computing time due to the bootstrap, results are computed and cached. This code is available for users to see how the example data and results were created.