Tutoring Model for Intelligent Tutoring Systems: Adaptive Instructional Scaffoldings Based on Experiences of the Learner Digital Twin

Abstract

This study investigates the impact of scaffolding provided based on learners’ digital twins within Intelligent Tutoring Systems (ITS) on participants’ performance and instructional support behaviors. Employing a quasi-experimental design, the research involved 104 university learners who were divided into experimental and control groups. The experimental group utilized an adaptive ITS that identified digital twins to provide personalized scaffolding, while the control group used an adaptable ITS where learners chose their support types. Data were collected through pre-tests, post-tests, and interaction logs, focusing on the effects of the instructional supports provided by the two different systems on academic performance. Statistical analyses, including paired-sample t-tests and ANCOVA, indicated significant improvements in the experimental group’s academic performance, with effect sizes showing a higher impact compared to the control group. Additionally, the adaptive system led to reduced behaviors related to receiving instructional support, suggesting that learner digital twin–based scaffolding may support learning outcomes by reducing the burden of learner self-selection and providing system-guided instructional support.