Examining Students' Learning Behaviors During the Problem-Solving Process in a Serious Game: A Prediction Study

• In Event: A Broad View of the Complexity of Learning With Technology

Abstract

This study examined whether students’ activities, captured by the log data, can be used to predict their learning performance in a 3D immersive serious game designed for middle school science in order to understand how the students solve a problem by interacting with various cognitive tools built in the environment. The log files were analyzed to build ordinal logistic regression models that predict students’ performance score over the four stages of the problem-solving process. Each model, at each stage, showed different positive and negative predictors of students’ performance over time and confirmed the successful learning behaviors. Findings also revealed emerging patterns in the students’ scientific inquiry processes.

Authors

• Jina Kang, The University of Texas - Austin

• Min Liu, The University of Texas - Austin