DASS-21 Analysis: Psychometric Properties and Anxiety Predictors

Overview

This repository contains the R Markdown script (DASS_analyses.Rmd) used for conducting a comprehensive psychometric analysis of the Depression, Anxiety, and Stress Scale-21 (DASS-21) and identifying factors influencing anxiety symptoms. The analyses were performed on data from undergraduate students at the Federal University of Ouro Preto, Brazil, as part of a larger research study on music therapy.

The script performs three main sets of analyses:

1. Data Preparation and Descriptive Statistics: Imports, cleans, and imputes data, followed by a summary of participant sociodemographics.
2. Internal Structure Validity and Reliability Analysis (DASS-21): Investigates the three-factor structure of the DASS-21 using Confirmatory Factor Analysis (CFA) and assesses the reliability of its subscales.
3. Predictors of Anxiety: Utilizes linear regression to determine which sociodemographic and lifestyle factors predict DASS-21 anxiety scores.
4. Network Psychometric Analysis: Explores the interrelationships among DASS-21 anxiety symptoms using network analysis.

File Structure

• DASS_analyses.Rmd: The main R Markdown file containing all R code, outputs, and narrative explanations for the analyses.
• comp_reliability.R: A custom R script containing a function to calculate composite reliability for the DASS-21 subscales. This script is sourced by DASS_analyses.Rmd and must be present in the same directory.
• Rendered Output: This repository might also contain a rendered version of the Rmd file DASS_analyses.pdf for easier viewing of the full report.

Prerequisites

To run the DASS_analyses.Rmd script successfully, you will need:

1. R: A recent version of R installed (developed with R version 4.4.2).
2. RStudio (Recommended): An IDE like RStudio facilitates working with R Markdown files.
3. R Packages: Install the following R packages:

   # Run these lines in your R console if you don't have the packages installed
   install.packages(c(
     "readxl",     # For reading Excel files
     "dplyr",      # For data manipulation
     "mice",       # For multiple imputation
     "psych",      # For descriptive statistics (like describe())
     "lavaan",     # For Confirmatory Factor Analysis (CFA)
     "semTools",   # For additional SEM tools (like reliability())
     "semPlot",    # For plotting SEM models
     "olsrr",      # For regression diagnostics and stepwise selection
     "car",        # For regression diagnostics (like Durbin-Watson)
     "lm.beta",    # For standardized beta coefficients in regression
     "bootnet",    # For network psychometric analysis
     "qgraph",     # For plotting networks (dependency of bootnet)
     "knitr",      # For creating dynamic reports
     "kableExtra", # For enhancing HTML/LaTeX tables
     "tidyr"       # For data tidying (e.g., pivot_longer)
   ))

4. Custom Reliability Script: The comp_reliability.R file must be in the same directory as DASS_analyses.Rmd.

Data

• This repository does not include the raw data file (~/dados_limpos.xlsx) due to privacy and data sharing restrictions.
• You must obtain this file separately and ensure it is accessible by the R script.
• Crucially, you MUST update the file path inside the DASS_analyses.Rmd script to point to the correct location of your data file:
  • Look for the line data <- read_excel("~/dados_limpos.xlsx") (or similar) and modify the path accordingly.

Usage

1. Install Prerequisites: Ensure R and all required R packages are installed (see list above).
2. Prepare Data: Place your data Excel file (e.g., dados_limpos.xlsx) in an accessible directory.
3. Place Custom Script: Ensure comp_reliability.R is in the same directory as DASS_analyses.Rmd.
4. Update File Paths: Open DASS_analyses.Rmd in RStudio or a text editor and modify the file path within the read_excel() function to match the location of your data file.
5. Run Analysis: Open DASS_analyses.Rmd in RStudio. You can:
   • Run individual code chunks sequentially.
   • Use the "Run" -> "Run All" command.
   • Knit the document (e.g., to HTML or PDF using the "Knit" button) which will execute all code chunks and generate a full report.

Analysis Sections in DASS_analyses.Rmd

The script is organized into the following main sections:

1. Setup and Package Loading: Loads all necessary R packages.
2. Data Import and Cleaning:
   • Reads the data from an Excel file.
   • Removes specified irrelevant columns.
   • Calculates and displays the proportions of different academic community members participating.
   • Subsets the data to include only undergraduate students for subsequent analyses.
3. Data Imputation:
   • Identifies a single missing value in the DASS-21 items for the undergraduate subgroup.
   • Performs multiple imputation using Predictive Mean Matching (PMM) via the mice package.
   • Uses the first imputed dataset for subsequent analyses.
4. Sociodemographic Data:
   • Presents descriptive statistics (mean, SD, median, range for age; frequencies and percentages for categorical variables) for the undergraduate student sample.
   • Includes age, ethnicity/race, biological sex, gender identity, religion, yoga/meditation practice, physical exercise, sleep satisfaction, and self-reported mental illness diagnosis.
5. Evidence of Internal Structure Validity and Reliability of the DASS-21:
   • Confirmatory Factor Analysis (CFA):
     • Specifies and fits the theoretical three-factor model (Anxiety, Stress, Depression) of the DASS-21 using lavaan.
     • Employs the WLSMV estimator, appropriate for ordinal item data.
     • Evaluates model fit using standard indices (χ², CFI, TLI, RMSEA).
     • Visualizes the factor model using semPlot.
   • Reliability Testing:
     • Calculates various reliability coefficients (Cronbach's Alpha, Ordinal Alpha, McDonald's Omega via semTools).
     • Calculates Composite Reliability using the custom comp_reliability.R script.
   • Factor Score Estimation: Estimates latent factor scores for Anxiety, Stress, and Depression using lavaan::lavPredict.
6. Testing Which Sociodemographical Factors Predict Anxiety:
   • Conducts a multiple linear regression with DASS-21 Anxiety factor scores as the outcome variable.
   • Uses stepwise forward selection (based on AIC) via olsrr to identify significant predictors among sociodemographic and lifestyle variables.
   • Presents the final model summary, ANOVA table, and parameter estimates (including standardized beta coefficients).
   • Checking Model Assumptions: Verifies residuals normality (Shapiro-Wilk test), homoscedasticity (Breusch-Pagan test), and residual autocorrelation (Durbin-Watson test).
7. Network Psychometric Analysis (DASS-21 Anxiety Items):
   • Estimates a network structure of the DASS-21 anxiety items using bootnet with the EBICglasso method.
   • Visualizes the network.
   • Calculates and plots centrality indices (Strength, Closeness, Betweenness, Expected Influence).
   • Assesses the stability of centrality indices using case-dropping bootstrapping and corStability.
8. Session Information: Includes output from sessionInfo() to document R and package versions.

Outputs

The DASS_analyses.Rmd script, when run or knitted, will produce:

• Descriptive statistics tables for sociodemographic data.
• CFA model fit statistics and parameter estimates.
• Reliability coefficient estimates.
• Linear regression model summaries and diagnostic test results.
• Network plots and centrality index tables/plots.
• Correlation stability analysis results for network centrality.
• All outputs are embedded within the rendered document if knitted (e.g., to HTML or PDF).

Important Notes

• File Paths: The most critical step for reproducibility is updating the path to your data file within DASS_analyses.Rmd.
• WLSMV Estimator: The CFA uses the WLSMV estimator, appropriate for ordinal DASS-21 data. Fit indices reported by lavaan for WLSMV are the correctly adjusted ones.
• comp_reliability.R: This custom script is essential for calculating composite reliability as presented in the report.

How to Cite

If you use or adapt this code or analysis structure, please cite this repository.

Citing this Repository/Code:

Pedrosa, F. (2025). DASS-21 Analysis: Psychometric Properties and Anxiety Predictors. GitHub Repository. https://github.com/FredPedrosa/DASS_21/

Author

• Frederico Pedrosa
• fredericopedrosa@ufmg.br

License

This project is licensed under a modified version of the GNU General Public License v3.0.  Commercial use is not permitted without explicit written permission from the author.