Developing Authoring Tools for Simulation-Based Intelligent Tutoring Systems: Lessons Learned

Abstract

Intelligent tutoring systems have a long history of significantly improving student performance. Unfortunately, they also have a long history of being very expensive to develop while producing very short shelf lives. To address these deficiencies, we developed and evaluated an authoring suite known as the Rapid Adaptive Coaching Environment (RACE). While developing RACE, we frequently encountered tension between maximizing the power of the intelligent tutoring system that authors would produce and minimizing the level of effort associated with producing this tool. In this paper, we will attempt to highlight this tension as we consider the five core design challenges that we addressed in the RACE prototype. We will review the formal and informal evaluations conducted by the development team and discuss how those evaluations led to the maturation of the design.