Analysis of the Effectiveness of Feedback Provision in Intelligent Tutoring Systems

The effectiveness of feedback provision in an Intelligent Tutoring System (ITS) is a crucial aspect to be considered when developing a tutoring system. The system is proven beneficial if it can act like a one-to-one human tutor, which provides feedback based on student learning aptitude and performance. Many researchers have developed ITSs with the intention of assisting students in teaching programming concepts, algorithms, and writing computer programs. However, there have been very few studies that have focused on reviewing these works. The studies are merely concerned with reviewing the characteristic, application, evaluation, and supplementary features of ITSs across different educational fields from 2007 until 2017. As a result, a comparative evaluation was conducted with the goal of analyzing the feedback provided by existing works in ITSs as well as the techniques used to develop a student model. The results of this study have indicated that Constraint-Based modeling, Model Tracing, Natural Language processing, and Deep Learning with delayed feedback are the most appropriate feedback for students learning to code. Bayesian Network is the most commonly used technique by researchers with immediate and delayed feedback to help students learn programming concepts and algorithms better.