projectgst: A Shiny Application for Generalizability Study Analysis in Nonformal Education Settings

Area of Science:

• Psychometrics and Educational Measurement
• Statistical Modeling
• Data Analysis Software

Background:

• Generalizability Theory (G-Theory) offers a robust framework for decomposing score variance into multiple error sources.
• Traditional G-Theory analyses demand specialized programming skills, creating a barrier for many applied researchers.
• Existing methods lack a unified, user-friendly interface for comprehensive G-Theory analysis.

Purpose of the Study:

• To introduce projectGST, an open-source R Shiny application designed for code-free Generalizability Theory analyses.
• To provide a unified platform integrating preliminary statistical analyses with G-Theory estimation and optimization.
• To enhance the accessibility of G-Theory for applied researchers across various educational contexts.

Main Methods:

• Development of an R Shiny application, projectGST, offering a graphical user interface for G-Theory.
• Integration of descriptive statistics, normality tests, correlation matrices, Cronbach's alpha, and intraclass correlation coefficients.
• Implementation of G-Study variance component estimation and D-Study reliability optimization for crossed and nested designs.

Main Results:

• projectGST provides a unified, code-free interface for conducting comprehensive G-Theory analyses.
• The application successfully integrates preliminary statistical analyses with G- and D-Study procedures.
• Demonstrated utility through simulated datasets in nonformal education assessment contexts.

Conclusions:

• projectGST significantly lowers the barrier to entry for applying Generalizability Theory in research.
• The application facilitates more accessible and efficient reliability estimation and optimization.
• projectGST empowers researchers to conduct advanced psychometric analyses without extensive programming knowledge.