An Effect Size Measure for Raju's Differential Functioning for Items and Tests.

Area of Science:

• Psychometrics
• Educational Measurement
• Item Response Theory

Background:

• Differential item functioning (DIF) is crucial for ensuring test fairness and validity.
• Existing effect size measures may not adequately capture noncompensatory DIF.
• Standardized reporting of DIF effects is needed for robust test analysis.

Purpose of the Study:

• To establish a novel effect size measure for noncompensatory differential item functioning (NCDIF).
• To provide a benchmark for identifying moderate and large NCDIF using the Mantel-Haenszel parameter.
• To offer guidance for reporting NCDIF in test items.

Main Methods:

• Developed an effect size measure for NCDIF based on the Mantel-Haenszel parameter.
• Investigated the influence of model, discrimination (a), and difficulty (b) parameters on NCDIF effect size.
• Generated tables detailing NCDIF effect size across various parameter levels (a, b, c).

Main Results:

• The proposed NCDIF effect size measure is sensitive to item parameters (a, b, c).
• Established specific effect size values for moderate and large NCDIF.
• Provided empirical data through tables illustrating NCDIF effect size variations.

Conclusions:

• The new NCDIF effect size measure offers a standardized approach to quantifying DIF.
• This measure facilitates the accurate identification and reporting of significant DIF in test items.
• Recommendations for interpreting moderate and large NCDIF are established, improving psychometric analysis.