The Impact and Detection of Uniform Differential Item Functioning for Continuous Item Response Models

Area of Science:

• Psychometrics
• Educational Measurement
• Statistical Modeling

Background:

• Item response theory (IRT) is widely used for categorical data.
• Continuous response models (CRMs) are increasingly used with computer-based testing.
• Investigating differential item functioning (DIF) in CRMs requires further research.

Purpose of the Study:

• To compare the effectiveness of three methods for assessing DIF in CRMs.
• To evaluate regression, MIMIC model, and factor invariance testing for DIF detection.
• To identify optimal methods for DIF analysis in continuous response data.

Main Methods:

• A simulation study was conducted to compare DIF detection methods.
• Methods assessed include regression, the MIMIC model, and factor invariance testing.
• Performance was evaluated based on Type I error control and statistical power.

Main Results:

• The MIMIC model exhibited effective Type I error control.
• The MIMIC model demonstrated relatively high power for detecting DIF.
• Factor invariance testing and regression showed limitations in performance.

Conclusions:

• The MIMIC model is a promising approach for DIF detection in CRMs.
• Findings inform the selection of appropriate methods for DIF analysis in continuous data.
• Further research can refine DIF detection strategies for advanced measurement models.