Does Answer Option Positioning Lead to DIF in Reading Comprehension? An Analysis with Parametric Logistic Item Response Theory Model Trees

• In Event: Virtual Poster Session
  In Virtual iPresentation Only: SIG-Rasch Measurement Virtual Poster Session

Abstract

This study examined whether response option positioning in a high-stakes reading test contributed to DIF across booklet versions and gender. Rasch tree analysis was applied to responses from 10,000 examinees on 20 items across four booklets using the psychotree package in R. The Rasch tree identified three nodes and flagged four items with large DIF. Booklet version, especially in interaction with gender, was a major contributor. Content analysis suggested that position effects (e.g., primacy, recency, or proximity between correct answers and distractors) may have influenced performance. This study highlights the utility of Rasch trees for detecting how subtle design features and subgroup characteristics interact to produce DIF and compromise test fairness.

Author

• Farshad Effatpanah Hesari, Technical University of Dortmund; Islamic Azad University, Mashhad Branch