Solving the measurement invariance anchor item problem in item response theory

Area of Science:

• Psychometrics
• Educational Measurement
• Statistical Modeling

Background:

• Measurement invariance testing is crucial for valid cross-group comparisons.
• A key challenge is selecting invariant anchor items, as some may exhibit differential item functioning (DIF).
• Existing methods for anchor item selection lack comparative efficacy data.

Purpose of the Study:

• To compare the performance of various anchor item selection strategies.
• To identify the most effective approach for selecting optimal anchor items in measurement invariance testing.
• To provide practical recommendations for improving DIF analyses.

Main Methods:

• A large-scale simulation study was conducted.
• Eleven variations across five different anchor item selection approaches were evaluated.
• The simulation assessed power and Type I error rates for each approach.

Main Results:

• A recently proposed 2-stage procedure by Lopez Rivas, Stark, and Chernyshenko (2009) demonstrated superior performance.
• This 2-stage approach consistently provided optimal statistical power.
• The procedure effectively maintained nominal Type I error rates across most simulated conditions.

Conclusions:

• The 2-stage anchor item selection method is highly recommended for practical invariance testing.
• This approach facilitates efficient and accurate identification of invariant anchor items.
• Implementing this method enhances the overall efficacy of differential item functioning (DIF) tests.