Abstract
In this paper we explore available tools for integrating intelligent tutoring communications in game-based learning platforms and examine theory-based techniques for delivering explicit feedback in such environments. The primary tool influencing the design of this research is the open-source Generalized Intelligent Framework for Tutoring (GIFT), a modular domain-independent architecture that provides the tools to author, deliver, and evaluate intelligent tutoring technologies. Influenced by social cognitive theory and cognitive load theory based research, the resulting experiment looks at varying approaches for using an Embodied Pedagogical Agent (EPA) defined as a tutor in a game-based training environment. Treatments were authored to examine tradeoffs between embedding an EPA directly in a game, embedding an EPA in GIFT’s browser-based Tutor-User Interface (TUI), or using audio prompts alone with no social grounding. A condition effectiveness evaluation is presented examining treatment effects on performance, reported mental demand, and perceived usefulness of feedback. Outcomes support the use of an EPA present in GIFT’s TUI when compared to effectiveness scores on audio prompts alone and the game embedded EPA.