Effects of Prior Knowledge in Mathematics on Learner-Interface Interactions in a Learning-by-Teaching Intelligent Tutoring System

Abstract

This study attempted to determine the influence of prior knowledge in mathematics of students on learner-interface interactions in a learning-by-teaching intelligent tutoring system. One hundred thirty-nine high school students answered a pretest (i.e., the prior knowledge in mathematics) and a posttest. In between the pretest and posttest, they used the SimStudent, an intelligent tutoring system that follows a teaching-by-learning paradigm. The intervention period lasted for three consecutive days with 1 hour session each. SimStudent captured learner-interface interactions, such as time spent tutoring, number of quizzes conducted, and number of hints requested. It was disclosed that prior knowledge in term identification was the only skill that had a consistent, positive, and significant influence on learner-interface interaction with a SimStudent. Thus, the null hypothesis stating that prior knowledge in mathematics does not significantly influence interaction of students with a simulated student was partially rejected. It was concluded that the students may demonstrate or omit a skill, depending on their prior knowledge on identifying the terms of equations and the next step in solving equations. Recommendations and directions for future studies were presented.