Using Students’ Learning Curves to Assess the Quality of Hints and Feedback in an Intelligent Tutoring System

Abstract

With traditional education facing challenges in delivering personalized instruction due to large student-to-teacher ratios, Intelligent Tutoring Systems offer a promising alternative by providing tailored instruction to individual learners. This study investigates the impact of hints provided by an Intelligent Tutoring System (ITS) on the learning of students. The learning of students can be plotted in a learning curve which shows the relation between the number of attempts or training opportunities and the number of errors while learning to evaluate the performance of the students. Two groups are formed: one including students who requested hints and another including students who did not request hints while solving the exercises. This division allows for a comparison of the impact of hint utilization on the learning curves. The created learning curves have increasing error rates with low R^2 scores, resulting in insufficient evidence to suggest that learning has occurred. Several remarkable observations are provided to find differences in both groups. However, further analyses are needed to support these findings.