Date of this Version
Cai, L. (2015). Examining sources of gender DIF using cross-classified multilevel IRT models (Master’s thesis). Available from DigitalCommons @ University of Nebraska-Lincoln.
A substantial amount of research has focused on the detection of differential item functioning (DIF) in the past. However, DIF detection and the estimation of DIF effect size do not explain why it occurs. Recent studies have investigated how or why DIF may occur. Improvements in DIF analysis models have made it possible to explore additional covariates as potential sources of DIF by measuring the extent to which these covariates account for variation in performance. The current study examines variability in math performance accounted for by gender, which is referred as gender DIF. This study then investigates how the presence of gender DIF is explained by both person predictors (i.e., opportunity to learn; OTL) and item characteristics (i.e., item format). A cross-classification multilevel IRT model framework is used to demonstrate the relationship among item difficulty, gender, OTL, and item format. Data come from three countries participating in an international study of pre-service math teachers, the Teacher Education and Development Study in Mathematics (TEDS-M).
Adviser: Anthony D. Albano