REBL Scale

Dr. Trisha Shrum, Chris Donovan

September 08, 2025

Introduction

This is the repository for the Repeated Environmental Behavior Latent (REBL) Scale project. It contains all the code to build the REBL scale from raw data through visualizations.

REBL Package

Installation

To install the rebl package:

if (!require('remotes')) install.packages('remotes')
remotes::install_github('tshrum/rebl')

Using the Package

Recoding Survey

library(rebl)

# Example data frame
str(raw_example)

# Identify REBL items with regex. This just produces a character vector with all
# the item names. It is not necessary if you already have this handy. Here is
# an example with the food items:
food_items <- id_rebl_items(raw_example, '^food', perl = TRUE)

# Here we use regex to take everything that does not start with 'res', because
# 'respondent_id' is our only non-REBL column. If items all start with 'rebl',
# for example, pattern would just be '^rebl*'
rebl_items <- id_rebl_items(raw_example, '^(?!res).*', perl = TRUE)

# Check function documentation for reference
?id_rebl_items

# Recode REBL items from Yes/No to 1/0
df <- recode_rebl(raw_example, rebl_items)

# To reverse code REBL items (where 0 is the PEB), first identify them.
# Just using all food items as an example
reversed_items <- id_reversed_rebl_items(rebl_items, '^food')

# Now recode the reversed items in the df
df <- reverse_code_rebl_items(df, reversed_items)

Getting REBL Scores

# Run Rasch Model
model <- get_rasch_model(df, 'respondent_id', rebl_items)

# Throws warning if some items are not found in df
model2 <- get_rasch_model(df, 'respondent_id', c(rebl_items, 'ffff', 'zzzz'))

# Throws error if no items are found in df
model3 <- get_rasch_model(df, 'respondent_id', c('ffff', 'zzzz'))

# Get a df of REBL scores, including person fit statistics
rebl_scores <- get_rebl_scores(model)
str(rebl_scores)

# Only REBL scores, no fits
rebl_only <- get_rebl_scores(model, include_fits = FALSE)
str(rebl_only)

File Structure

The project is halfway through transitioning into a package (2025-09-03). The package has the usual structure:

• R/: Functions
• data/: Built-in datasets. Used in examples and testing.
• data-raw/: Scripts to create built-in datasets.
• man/: Documentation
• tests/: Test suite using testthat
• DESCRIPTION: Package information, license, author
• NAMESPACE: Exports and imports

In the meantime, functions will be duplicated across 3_functions/ and R/ until things are settled, at which point I hope to clear this up and create a more focused, functional, and portable workflow.

REBL Project

Clone Repository

Dependencies

The project is set up with renv. To restore the packages used in the project, use:

# install.packages('renv')
renv::restore()

This will install packages and versions as described in the renv.lock file.

Navigating the Project

After opening reb_scale.Rproj, start with table_of_contents.R. This will list all the relevant scripts in the order they should be run. Run the housekeeping.R script, which will load helper functions that are used throughout the project. Additional required packages are then included in the top of each script. Using ctrl/cmd + left click on the text of a file path in the table of contents will open that script in a new tab.

All the objects and packages required for each script are loaded at the top of the page, and all important outputs are saved at the end of each script. So, you can run the entire project from the table of contents, but you can also just skip to any script and run it from there.

Project File Structure

• 1_raw/ is for raw, untouched, datasets. Metadata is embedded into these files using # as a comment symbol.
• 2_clean/ is for cleaned (wrangled, recoded, imputed, etc.) data.
• 3_functions/ is where custom functions are stored. The housekeeping script will load these.
• 4_scripts/ is for all R scripts.
• 5_objects/ is for intermediate .rds objects.
• 6_outputs/ is where tabular outputs are kept.
• 7_plots/ is for figures, diagrams, pngs, pdfs, etc.