A number of studies have shown that intelligent tutoring systems can lead to enhanced learning outcomes in areas of mathematics, the sciences, programming, language acquisition, and others. However, the formal schooling system cannot possibly prepare people for all the skills and knowledge needs they will ever have during their lives. It is important therefore that the educational system helps people develop a general ability to get up to speed quickly in new domains. In order to do that students need to be able to manage their learning, for example, by setting goals, planning their learning, monitoring their progress, and responding appropriately to difficulties and errors.

These general learning skills are often being referred to as metacognition, or self-regulated learning (RSL). Bransford et al. (2000) suggest focusing on metacognition as one of three principles that should be applied to educational research and design, as stated in the influential volume “How People Learn.” A similar recommendation is given also in Clark and Mayer's (2003) book about e-learning design principles. In spite of these recommendations, it is still unclear (both in the traditional classroom and in tutoring systems) what exactly are appropriate goals for metacognitive tutoring and how these goals could be achieved. On one hand, educational programs focused on improving people’s general learning skills have not met with great success (e.g., Resnick 1987). On the other hand, the educational psychology literature presents a number of highly successful instructional programs focused on a more specific notion of meta-cognitive skills. For example, Carver and Mayer (1998) demonstrated that a well-designed 30 minutes instruction on debugging can improve students’ ability to identify and correct errors also in far-transfer tasks such as accuracy of general instructions, which are not related to programming. White and Frederiksen fostered learning, especially for low-achieving students, using a metacognitive cycle of inquiry, reflection, and generalization (White and Frederiksen, 1998).  

A key question is whether instructional technology can be as effective in fostering metacognitive skills as it is in teaching domain-specific skills and knowledge. On the face of it, the answer is positive. Currently, several tutoring systems support metacognition. For example, Conati and VanLehn (2000) and Aleven and Koedinger (2002) support self-explanation in order to promote learning of Physics and Geometry respectively. Roll et al (2006) attempt to teach students better help-seeking skills. Supporting metacognition seems especially important in inquiry, discovery, and hypermedia environments, in which much of the responsibility of the learning process is put on the learner. De Jong and van Joolingen (1998) describe ways in which open environments can support students’ metacognitive skills, for example, by supporting self-regulatory activities. Azevedo (2005) further details similar aspects of SRL. These include, for example, student control over setting subgoals and the use of learning resources. However, it remains largely unknown exactly how Intelligent Tutoring Systems can help students acquire better metacognitive skills and thereby become better learners with respect to domain-specific skills and knowledge.

This workshop will attempt to identify means and improve our understanding with regard to teaching metacognition using Intelligent Tutoring Systems.

 

References