Using Machine Learning Models to Predict Normalized Learning Gains of High-School Students Playing a Simulation Game-Based Learning Environment

• In Event: AI-Mediated Learning Environments: Scaffolding Inquiry, Metacognition, and Engagement

Abstract

This study reports the results of a newly developed simulation, game-based learning environment (SGBLE) that teaches learners about the spread of viruses, called Out-break Simulator. This study reported on 56 high school students as they played this SGBLE, and used a supervised machine learning model using self-reported metacognitive regulation and cognitive load to predict their learning gains while playing this game. A random forest regression model indicated that participants’ learning gains about virology were predicted by a combination of metacognitive and cognitive load factors, most notably, reported mental demands of the task. These findings provide evidence of the role that cognitive load indicators play in the prediction of performance and metacognitive regulation of learners immersed in simulation-based learning.

Authors

• Matthew Moreno, University of Central Florida

• Megan Wiedbusch, University of Central Florida

• Barrie Robinson, University of Idaho

• Terrance Soule

• Justin Riggs, University of Idaho

• Roger Azevedo, University of Central Florida