Anchor Point Selection: Scale Alignment Based on an Inequality Criterion.

Area of Science:

• Psychometrics
• Educational Measurement
• Item Response Theory

Background:

• Detecting differential item functioning (DIF) is crucial for fair test comparisons across groups.
• Current Rasch model approaches rely on anchor items selected via statistical tests or heuristics.
• Aligning item parameters across groups is necessary for accurate DIF detection.

Purpose of the Study:

• To propose and evaluate the Gini Index as an alternative criterion for selecting anchor points in DIF analysis.
• To compare the Gini Index method with the established alignment approach.
• To explore the additional information provided by the Gini Index criterion plot for identifying multidimensionality.

Main Methods:

• Applied the Gini Index, an economic inequality measure, to optimize the overlap of item parameter estimates between groups.
• Conducted extensive simulation studies and analyzed empirical data.
• Compared the Gini Index criterion with the alignment approach by Asparoughov and Muthén.

Main Results:

• The Gini Index effectively identifies an optimal anchor point for aligning item parameters.
• The criterion plot generated by the Gini Index offers insights into potential multidimensionality.
• The Gini Index approach demonstrates comparable or superior performance to the alignment method in simulations.

Conclusions:

• The Gini Index provides a robust and informative method for anchor point selection in DIF analysis within the Rasch model.
• This approach enhances the detection of DIF and can aid in uncovering sources of measurement bias.
• The method is computationally efficient and extendable to complex, multi-group scenarios.