Evaluating the Impact of Interface Agents in an Intelligent Tutoring Systems Authoring Tool

This paper describes the evaluation of a speech-driven cartoon agent that has been embodied in the interface of a Web-based Intelligent Tutoring Systems (ITSs) authoring tool. The agent is mainly used in the student interface of the ITSs that are produced by the authoring tool and represents either a student or an instructor. The agent was evaluated during an experiment in terms of the effect that it could have on students’ interaction with the generated ITSs. The participants in the experiment were divided into three groups: one third of them worked with a version of the ITS in which the agent was supposed to be an instructor, another third worked with a version of the ITS in which the agent was acting as a co-student and the rest worked with an agent-less version. The results from this study seem to confirm the hypothesis that a pedagogical agent incorporated in an ITS can improve students’ experience. On the other hand, the hypothesis that the presence of the agent significantly enhances learning was rejected based on our data. Comparing the two agent versions of the co-student and instructor, the most important finding was that students seem to feel more comfortable working with a virtual co-student than with a virtual instructor.