A multiverse analysis of the associations between internet use and well-being

• Matti Vuorre
• Andy Przybylski

This repository contains the code and simulated dataset associated with "A multiverse analysis of the associations between internet use and well-being" (Vuorre & Przybylski).

• preprint: https://doi.org/10.31234/osf.io/jp5nd
• repository: https://github.com/digital-wellbeing/gwp-multiverse
• archive: https://doi.org/10.5281/zenodo.7774923

Reproduce

Environment

We coded the analyses in R, and used renv to manage the project's R environment. To ensure that you run computations in the same environment, first restore it with renv::restore(prompt = FALSE).

Environment variables

Use the variables in .Renviron to e.g. point to a local GWP data file, or use a number of CPUs more appropriate to your machine.

• MAX_CORES
  • Number of CPUs to use.
• DATA_PATH_GWP
  • Path to the actual GWP data.
• DATA_PATH_GWP_CLEAN
  • Path to cleaned GWP data
• DATA_PATH_SYN
  • Location of the synthetic data file.

Data

If you do not have access to the GWP data, we provide a small simulated dataset for reproducibility. If you do have access, modify environment variables (above) accordingly.

Run computations

All the code needed for our analyses is in ms.Rmd, an R Markdown file that can be "rendered" to a PDF. You can either run code in that file manually, or rmarkdown::render('ms.Rmd') to produce the PDF. Note however that you'll need a working LaTeX installation for the latter.

Help

If all else fails, please email Matti at mjvuorre@uvt.nl.