Power Formulas for the Logistic Regression Tests for Uniform and Nonuniform Differential Item Functioning

• In Event: 72.021 - Topics in Differential Item Functioning

Abstract

In this paper we derive a power formula of the Logistic Regression (LR) tests for the differential item functioning (DIF). Logistic regression is a popular method for detecting uniform and non-uniform differential item functioning (DIF) effects. As shown in the derived formula, the asymptotic power is related to the sample size, effect size of DIF, the item response function (IRF) and the distribution of the latent trait of the reference and the focal groups. A Monte Carlo simulation study shows that the theoretical values calculated from the formulas derived in the paper are close to what are observed in the simulated data.

Authors

• Zhushan Mandy Li, Boston College

• Jing Jiang, Boston College