Multidimensional Sequential Response Model: Assessing Multiple Abilities in Process Data Analysis

• In Event: Item Response Theory

Abstract

Expanding on the sequential response model (SRM) (Han et al., 2021), this study proposes multidimensional sequential response models (MSRM) that can estimate multidimensional latent constructs from dynamic response processes in computer-based assessments. Monte-Carlo simulation experiments shows that that the MSRM performs well in parameter recovery and fits better than the unidimensional SRM when the underlying latent ability is of multidimensional. We also showcase the use of MSRM using a real-world interactive assessment task.

Authors

• Yuting Han, Beijing Language and Culture University
  Presenting Author

• Feng Ji, University of California - Berkeley
  Non-Presenting Author

• Hongyun Liu, Beijing Normal University
  Non-Presenting Author