Using Process Data in Diagnostic Classification Modeling with More Flexible Distributional Assumptions

• In Event: Electronic Board Session I

Abstract

Jointly modeling of process data and item responses helps improve classification accuracy and provides more comprehensive diagnosis in cognitive diagnostic modeling with the premise of correct model specifications. This study proposes the use of the flexible Conway-Maxwell-Poisson (CMP) distribution for action counts in this context for more accurate statistical inference.

Author

• Xin Qiao