Educational Game and Intelligent Tutoring System

Abstract

Educational games and intelligent tutoring systems (ITS) both support learning by doing, although often in different ways. The current classroom experiment compared a popular commercial game for equation solving, DragonBox and a research-based ITS, Lynnette with respect to desirable educational outcomes. The 190 participating 7th and 8th grade students were randomly assigned to work with either system for 5 class periods. We measured out-of-system transfer of learning with a paper and pencil pre- and post-test of students’ equation-solving skill. We measured enjoyment and accuracy of self-assessment with a questionnaire. The students who used DragonBox solved many more problems and enjoyed the experience more, but the students who used Lynnette performed significantly better on the post-test. Our analysis of the design features of both systems suggests possible explanations and spurs ideas for how the strengths of the two systems might be combined. The study shows that intuitions about what works, educationally, can be fallible. Therefore, there is no substitute for rigorous empirical evaluation of educational technologies.