Reproducible Analysis Script

These scripts allow the reproduction of all the data analysis, tables and figures used in the paper, including the meta-analysis of all the results.

The only caveat is that the 95% confidence intervals presented in the paper may not match perfectly the ones produced by this script. We forgot to set a reproducible seed for the random number generator in our original script, making those results not 100% reproducible. The output of these scripts, on the other hand, should always be the same.

In order to run the full reproducible analysis, open the Rproject file on a RStudio session and source the first script "00-ReproduceAnalyses-Tucker_et_al_2021.R", which sequentially run the individual experimental scripts. The requisite data is automatically downloaded from figshare and need not be downloaded manually or independently.

These scripts were build on R version 4.0.3, on a macOS Mojave 10.14.6, using RStudio version 1.3.1093. For more information about all packages and their respective versions that were used to successfully run these scripts, check out the file "tucker2021_original_session_info_for_comparison.txt" available here.

There is GitHub research compendium for the same scripts at https://github.com/diogo-almeida/TuckerIdrissiAlmeida2021. It contains an experimental feature of being able to run an RStudio session on the browser with all the correct package versions and R version, so users do not need to worry about Rstudio and R version conflicts in their own local environments.