A tutorial on Bayesian item response theory: An illustration using the Teacher Stress Inventory-Short Form.

Area of Science:

• Educational Psychology
• Psychometrics
• Quantitative Psychology

Background:

• Item Response Theory (IRT) is a standard psychometric tool for educational assessments.
• Frequentist IRT methods have limitations regarding assumptions and computational power.
• Bayesian approaches offer a flexible and robust alternative for IRT analyses.

Purpose of the Study:

• To introduce school psychology researchers to Bayesian methods for Item Response Theory (IRT) analyses.
• To highlight the advantages of Bayesian IRT over traditional frequentist approaches.
• To demonstrate a comprehensive Bayesian workflow for IRT.

Main Methods:

• Overview of the Bayesian IRT workflow: model specification, prior selection, predictive checks, model estimation, posterior interpretation, and sensitivity analyses.
• Application of Bayesian IRT to the 16-item Teacher Stress Inventory-Short Form (TSI-SF).
• Utilized a sample of 329 teachers for the analysis.

Main Results:

• The Bayesian IRT analysis indicated that the TSI-SF is best modeled by a three-correlated-traits structure.
• The identified traits are Discipline and Motivation, Professional Investment, and Work-Related Stress.
• This provides a nuanced understanding of the sources of teacher stress.

Conclusions:

• Bayesian IRT provides a powerful framework for analyzing educational assessment data, offering flexibility and robust inference.
• The findings support a multidimensional structure of teacher stress as measured by the TSI-SF.
• School psychology researchers can benefit from adopting Bayesian methods for complex psychometric modeling.