In today's society, people constantly face the challenge of acquiring new skills and knowledge. Such learning tasks are easier for those who are able to use available sources of help judiciously and who can accurately assess their own knowledge. The proposed research will test the hypothesis that computer-based interactive learning environments can help people become better at metacognitive skills such as help seeking and self-assessment. Building on successful work in Cognitive Tutors, the project will involve the creation and evaluation of a computer tutor that guides students at the metacognitive level, as they solve problems. The tutor will be equipped with a detailed model of help-seeking skills. A series of controlled experiments will be conducted to evaluate the added value of metacognitive tutoring. The proposed research addresses open questions in the fields of intelligent tutoring systems, cognitive science, and educational psychology. The proposed research will make it clear whether intelligent tutoring systems are suitable vehicles for helping students acquire metacognitive skills. It will produce models of help seeking that are more detailed than those described in the literature and, if successful, will demonstrate instructional methods that improve students' skill in seeking help and self-assessment and thereby their ability to learn.
