Parameter recovery for the rating scale model using PARSCALE.

Area of Science:

• Psychometrics
• Educational Measurement
• Statistical Modeling

Background:

• Item response theory (IRT) models are crucial for educational and psychological assessments.
• Andrich's rating scale model (RSM) is a widely used polytomous IRT model.
• Understanding parameter recovery is essential for model validity.

Purpose of the Study:

• To evaluate item and trait parameter recovery for Andrich's rating scale model.
• To investigate the influence of data characteristics on model performance.
• To compare calibration sample size requirements with other polytomous IRT models.

Main Methods:

• Simulated data matrices were generated varying item scale distribution, response categories, trait level distribution, and sample size.
• The PARSCALE computer program was used for analysis.
• 400 data matrices were analyzed across 10 replications per condition.

Main Results:

• Accurate item and trait parameter estimates were achieved across all tested sample sizes.
• Sample size impacted item parameter recovery but not trait parameter recovery.
• The distribution of trait levels had minimal effect on item parameter recovery.

Conclusions:

• Andrich's rating scale model demonstrates robust parameter recovery.
• Smaller calibration sample sizes may be sufficient for Andrich's RSM compared to other polytomous IRT models.
• The model's efficiency makes it suitable for various assessment contexts.