Test Item Bias Analysis Using Differential Item Functioning (DIF): A Mantel-Haenszel Chi-Square Statistics Approach



Author Information

Arlene Nisperos Mendoza, Pangasinan State University, Philippines

Abstract

This study highlights the importance of implementing Differential Item Functioning (DIF) analysis to assess the fairness and validity of educational measures. The analysis examines possible test item biases against certain groups of test-takers based on factors like age, sex, socio-economic status, and school type. Utilizing the Mantel-Haenszel Chi-Square Statistic, the study identified biased test items, with over one-third exhibiting bias, consequently compromising the assessment's fairness and validity. The findings demonstrated that age, sex, socioeconomic status, and the type of educational institution exerted a discernible influence on the disparities observed in students' performance on the examination. Moreover, it was ascertained that age played a particularly significant role in these variations. Removing potentially biased items resulted in a more equitable and valid assessment, emphasizing the importance of identifying potential biases to enhance the test's quality and reliability, ultimately contributing to the improvement of educational assessment.


Paper Information

Conference: PCE2025
Stream: Assessment Theories & Methodologies

This paper is part of the PCE2025 Conference Proceedings (View)
Full Paper
View / Download the full paper in a new tab/window


To cite this article:
Mendoza A. (2025) Test Item Bias Analysis Using Differential Item Functioning (DIF): A Mantel-Haenszel Chi-Square Statistics Approach ISSN: 2758-0962 The Paris Conference on Education 2025: Official Conference Proceedings (pp. 707-724) https://doi.org/10.22492/issn.2758-0962.2025.54
To link to this article: https://doi.org/10.22492/issn.2758-0962.2025.54


Comments & Feedback

Place a comment using your LinkedIn profile

Comments

Share on activity feed

Powered by WP LinkPress

Share this Research

Posted by James Alexander Gordon