Differential Item Functioning Detection and Effect Size: A Comparison between Logistic Regression and Mantel-Haenszel Procedures

This article compares several procedures in their efficacy for detecting differential item functioning (DIF): logistic regression analysis, the Mantel-Haenszel (MH) procedure, and the modified Mantel-Haenszel procedure by Mazor, Clauser, and Hambleton. It also compares the effect size measures that...

Full description

Saved in:
Bibliographic Details
Published in:Educational and psychological measurement 2004-12, Vol.64 (6), p.903-915
Main Authors: Hidalgo, M. Dolores, LÓPez-Pina, JosÉ Antonio
Format: Article
Language:English
Subjects:
Biological and medical sciences
Comparative analysis
Effect Size
Evaluation Methods
Evaluation Research
Fundamental and applied biological sciences. Psychology
Item Difficulty
Measurement
Psychological Studies
Psychology
Psychology. Psychoanalysis. Psychiatry
Psychology. Psychophysiology
Psychometrics
Psychometrics. Statistics. Methodology
Regression (Statistics)
Regression analysis
Statistical Analysis
Statistics. Mathematics
Studies
Test Bias
Test Items
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This article compares several procedures in their efficacy for detecting differential item functioning (DIF): logistic regression analysis, the Mantel-Haenszel (MH) procedure, and the modified Mantel-Haenszel procedure by Mazor, Clauser, and Hambleton. It also compares the effect size measures that these procedures provide. In this study, different conditions of item parameters (difficulty and discrimination) and DIF magnitude were manipulated. Furthermore, both uniform and nonuniform DIF conditions were simulated. Results suggest that logistic regression analysis generally detected more items with DIF than the standard MH procedure and the modified MH procedure for symmetrical nonuniform DIF. The DIF effect size measures based on logistic regression, however, appeared to be insensitive to the specified DIF conditions.
ISSN:0013-1644
1552-3888
DOI:10.1177/0013164403261769