Mapping the Perception-space of Facial Expressions in the Era of Face Masks

This repository contains all data and scripts to reproduce statistical analysis, figures and tables from the paper “Mapping the Perception-space of Facial Expressions in the Era of Face Masks” by Verroca, de Rienzo, Gambarota and Sessa (2022). The project is also on Open Science Framework (https://osf.io/e2kcw/). Supplementary materials are available here

Repository

The repository is organized with the following structure:

• data/: contains the cleaned data. The only pre-processing step is to combine different files from the Gorilla platform within a single dataset for each participant.
  • dat_clean.rds: dataset with all participants, catch and valid trials
  • dat_fit: dataset without the neutral facial expression and extra pre-processing steps used for model fitting
  • cleaned/catch/: contains data from catch trials used as attention check during the task
    • dat_catch_ang: dataset with catch trials and all pre-processing steps
    • dat_catch_ang_acc: dataset with catch trials and all pre-processing steps with computed accuracy
  • cleaned/valid: contains data from valid trials used for plotting and modelling
    • dat_valid_ang: contains all participants, valid trials and all pre-processing steps
    • dat_valid_ang_final: contains only good participants (excluded from catch trials) with all pre-processing steps
• figures/: contains all figures included in the paper or supplementary materials
• tables/: contains all tables included in the paper or supplementary materials
• objects/: contains all R objects used through the project. In particular contains all post-processed fitted models.
• scripts/: contains all scripts to produce datasets, models, figures and tables. The numbering suggest the order to correctly reproduce the analysis.
• files/: contains extra files used in the project
• docs/: contains the Rmd script and all files to reproduce the supplementary materials document
• R/: contains all custom functions used through the project.

Model fitting

All models are computed within a Bayesian framework using the brms R package. Models were fitted using cluster computing in order to speed-up the process. The final size of each model is on average more than ~200mb and for this reason we included only post-processing information (within the objects/ folder, created with 05a/b_post_processing_*.R scripts). Running again the 04a/b_*_models.R scripts will reproduce the same results.

Packages

• rmarkdown
• bookdown
• knitr
• devtools
• dplyr
• flextable
• ggplot2
• here
• kableExtra
• magrittr
• tidyr
• cowplot
• forcats
• ggh4x
• ggpubr
• magick
• stringr
• cli
• renv
• rlang
• tools
• tidyverse
• brms
• tidybayes
• CircStats
• ftExtra
• officer
• purrr
• latex2exp
• ragg