Examining Measurement Invariance and Differential Item Functioning With Discrete Latent Construct Indicators: A Note...

Area of Science:

• Psychometrics
• Statistical Modeling

Background:

• Measurement invariance is crucial for valid latent construct evaluation.
• Existing methods may have limitations in detecting specific types of invariance violations with binary data.

Purpose of the Study:

• To present an extended latent variable modeling method for assessing measurement invariance.
• To enable precise localization of measurement invariance violations in loading or threshold parameters.
• To facilitate the study of differential item functioning (DIF) with binary items.

Main Methods:

• Extends the continuous indicator procedure using latent variable modeling.
• Incorporates the false discovery rate (FDR) approach for multiple testing.
• Applies to multiple binary or dichotomously scored items without guessing.

Main Results:

• The method effectively identifies violations in measurement invariance.
• It pinpoints specific parameters (loadings or thresholds) where violations occur.
• Demonstrates applicability without requiring a reference observed variable.

Conclusions:

• The proposed method offers a robust approach to measurement invariance testing for binary data.
• It provides a valuable tool for researchers evaluating latent constructs and item functioning.
• The technique is directly applicable to one- and two-parameter item response models.