Simulation Code for “Causal inference targeting a concentration index for studies of health inequalities”

This repository contains the R code used to reproduce the simulation results presented in the paper Ghasempour, M., de Luna, X. and Gustafsson, P.E. (2024).“Causal inference targeting a concentration index for studies of health inequalities.”
The simulation evaluates the performance of the proposed estimators under different model specifications.

---

Contents

The repository includes the following R scripts:

1. sim_par_job.R – Main simulation script. Runs the simulations for a given sample size and saves the output files.
2. contrast_sim_utility.R – Contains supporting functions required by the main simulation script.
3. post_processing.R – Processes and summarizes the simulation results into separate CSV files.

---

Instructions for use

1. Ensure that the files sim_par_job.R and contrast_sim_utility.R are located in the same directory.

2. Open R (or RStudio) and set the working directory to the folder containing these files:

   setwd("path/to/your/folder")

3. To run the simulation, open the file sim_par_job.R. Inside the script, specify:

   path <- "path/to/save/results/"
   n <- 1000
   r <- 1000

   Then run:

   source("sim_par_job.R")

   The simulation results will be generated and stored in the specified directory.

4. When the simulation has finished, open the post-processing script and specify the same sample size n and the number of replications r used in the simulation.

   n <- 1000
   r <- 1000

   After setting n and r, run:

   source("post_processing.R")

   This script creates separate CSV files containing the summarized results for:

   • G(0)

   • G(1)

   • G(2)

   • and the contrasts between them.

   The resulting CSV files correspond to the tables reported in the paper.

Additional notes

• All required R packages should be installed prior to running the code. The list of packages is provided at the beginning of each script.

• The results are automatically saved in the directory specified by the user in the path variable.

Citation

If you use this code, please cite the associated paper:

Ghasempour, M., de Luna, X., & Gustafsson, P. E. (2024).
Causal inference targeting a concentration index for studies of health inequalities.
arXiv:2410.08849