Using item response theory models to evaluate the Practice Environment Scale.

Area of Science:

• Psychometrics
• Educational Measurement
• Health Services Research

Background:

• Item response theory (IRT) provides advanced methods for analyzing measurement instruments.
• Evaluating the Practice Environment Scale (PES) is crucial for understanding healthcare settings.

Purpose of the Study:

• Introduce various item response theory models.
• Demonstrate the utility of IRT models by evaluating the Practice Environment Scale.
• Identify optimal IRT models for scale analysis.

Main Methods:

• Applied several IRT models: constrained and unconstrained graded response, partial credit, Rasch, and one-parameter logistic.
• Utilized Akaike information criterion (AIC) and Bayesian information criterion (BIC) for model selection.
• Assessed the performance of individual items within the Practice Environment Scale.

Main Results:

• The unconstrained graded response model and partial credit model demonstrated the best fit to the data.
• Most items in the Practice Environment Scale exhibited good performance characteristics.
• Identified specific items that could potentially be removed with minimal impact on the overall construct measurement.

Conclusions:

• The Practice Environment Scale is largely effective in measuring its intended construct.
• Refinement of the instrument is possible by removing redundant items.
• The scale's performance may vary across different administrative contexts or unit types.