Exploring Differential Facet Functioning Models

• In Event: Comparing, Decomposing, and Exploring Models

Abstract

Differential facet functioning (DFF) is an alternative to Differential item functioning (DIF) that can measure and is defined as the operationalization of the detection of measurement bias for interrelated and integrated group of items. This study investigated measurement invariance of students classified as ELL, exited ELL, and non-ELL students across content facets of the 2011 NAEP 8th grade science exam by comparing four models: the measurement invariant Rasch model (Rasch, 1960), a DIF model, and two DFF models. The study found that, out of the four models, the latent regression model fit the data the best, however the cross-classified MMM also showed DFF in a way that would complement the DIF analysis.

Authors

• Luke Stanke, University of Minnesota

• Mark L. Davison, University of Minnesota