An Application of Multilevel Cross-Classified Random Effects Model to Analyze Peer-/Self-Ratings in an Online Collaborative Learning Environment

• In Event: Roundtable Session 97
  In Roundtable Session: Multilevel Modeling Methods and Applications III

Abstract

With the popularity of collaborative learning, researchers suggested that peer and self assessment is a way to assess individual performance in collaborative group projects. Given the nature of collaborative learning, the dependency among group members makes multi-level statistical models suitable to analyze the score data collected from group learning contexts. However, the network data structure of peer/self assessment scores, which does not entail a pure-clustering data structure, adds a challenge to model peer/self assessment score data with conventional hierarchical linear models.
In this project, a cross-classified random effects modeling (CCREM) technique was applied to analyze peer/self assessment scores with network data structure to address the challenge and to model group effect, rater effect, and ratee effect.

Author

• Yueh-hui Vanessa Chiang, Stanford University