Code for On Exact Randomization-based Covariate-adjusted Confidence Intervals

The simulation study and data analysis for this paper was conducted in R version 4.0.5. You will need to run the following lines of code to install the following packages to replicate the study:

package_vec <- c('tidyverse', 'here', 'Rfast', 'matrixStats', 'fpCompare', 'ri2', 'kableExtra')
install.packages(package_vec)

For the simulation study, you need to do the following:

1. Run generate_data_sims.r
2. On a cluster run simulation_study.sh, which will run simulation_study.R (the main script to perform the simulations) in parallel. You can also re-write the code in simulation_study.R so that it can run without a shell script
3. Run collect_sim_output.R
4. Run analyze_sim_results.R

The data analysis can be performed by running analyze_phase_I_trial.R. The data is already is in the data folder, and was downloaded from the Supporting Information from Pair-switching rerandomization.