Detecting Differential Options Functioning Using Multinomial Logistic Regression in SPSS

• In Event: 52.083 - Roundtable Session 19
  In Roundtable Session: 52.083-19 - Advances in Differential Item Functioning Research

Abstract

Compared to differential item functioning (DIF), there have been few studies on how options of multiple-choice items function differently for distinct subgroups. Although some researchers have discussed the study of differential option functioning, the purposes and utilities of previous research were limited to a post hoc analysis for investigating the causes of DIF. Suggested terminologies and methods in previous research were not integrated, either. The authors newly defined the term differential options functioning (DOF) as a generic and neutral term. Also, we proposed a straightforward method for studying DOF based on multinomial logistic regression as a stand-alone, self-sufficient investigation rather than a post hoc analysis. The method was demonstrated with PIRLS 2011 data in SPSS.

Authors

• Minjeong Park, The University of British Columbia

• Amery D. Wu, University of British Columbia