Lasso Regularization Versus Tree-Based Modeling Approach for Differential Item Functioning Detection in Polytomous Items

• In Event: Marriott Roundtable Session One
  In Roundtable Session: Differential Item Functioning (Table 1)

Abstract

The current study evaluates the effectiveness of the new DIF detection approaches (Lasso Regularization, Tree-Based Modeling approach) with the traditional DIF detection approach (Lord's Chi-Squared) in polytomous items. The simulation results show the new DIF detection approaches are effective in DIF detection. And they might have great implications for psychometric research practice.

Author

• Wei Xu, The American Board of Anesthesiology
  Presenting Author