Quantifying Engagement Level in Gamification-Based Intelligent Tutors Using Behavioral and Cognitive Metrics

Abstract

Gamification-based intelligent tutoring systems are popular because of improved and increased motivation, engagement, performance, and learning. Among the various components of a successful gamified learning environment, the accurate evaluation of learner engagement plays a critical role in delivering adaptive and personalized learning experiences. Despite the growing interest in gamification, limited work has been done to quantitatively assess engagement using foundational interaction metrics. This study proposed a metric for evaluating the engagement level of a learner in a game. The proposed metric incorporates both behavioral and cognitive engagement factors, including response time, action accuracy, level completion rate, and difficulty handling. These attributes are mathematically combined using a weighted scoring model to classify learners into three distinct engagement levels: actively disengaged, not engaged, and actively engaged. The proposed metric is validated using a personalized small dataset of a programming learning game. The proposed metric evaluated the engagement level with 83.33% accuracy, demonstrating its potential for use in real-world intelligent tutoring environments. The dataset can be extended in the future, refining the metric using dynamic weights and integrating the model into large-scale gamified educational platforms.