| This thesis explores the design and evaluation of a collaborative, inquiry learning Intelligent Tutoring System for ill-defined problem spaces. The common ground in the fields of Artificial Intelligence in Education and Computer-Supported Collaborative Learning is investigated to identify ways in which tutoring systems can employ both automated coaching and collaborative techniques to support students as they learn. The resulting system, Rashi, offers feedback on student work by using an Expert Knowledge Base to recognize students' work. Evaluation in actual classrooms demonstrated that collaboration significantly improves students' contributions, and some evidence suggests that there is a significant positive correlation between the amount of coaching received and metrics that represent positive inquiry behavior. Finally, this thesis highlights the potential for combining coaching and collaboration such that 1) collaborative work can create more opportunity to provide automated coaching and 2) automated coaching can identify key moments when collaboration should be encouraged. |
