Using Firth’s Penalized Maximum Likelihood Estimation for Logistic Regression to Detect Polytomous Differential Item Functioning

• In Event: AERA Roundtable Session Sunday 11:40 am Four Seasons Ballroom 1
  In Roundtable Session: Advances in Differential Item Functioning (DIF) Detection: Methods, Applications, and Implications (Table 12)

Abstract

Logistic regression is a common method for differential item functioning (DIF) detection and is typically estimated by maximum likelihood estimation (ML), despite the fact that ML may produce biased estimates in situations of rare event data and small sample sizes. Firth’s penalization method (PML) may address this issue. The current study compared PML with ML estimation in logistic regression for polytomous DIF detection, focusing on rare event and small sample. The differences in power and type I error rates between PML and ML were negligible in the simulated conditions. We did not observe inflated type I error rates across the simulated conditions. We provide practitioners with sample size suggestions across various item difficulty averages for the concerns of power.

Authors

• Yue Xu, University of Florida

• Corinne Huggins-Manley, University of Florida

• M. David Miller, University of Florida