Abstract :
Feedback on students’ actions is the major form of improving students’ performance in Intelligent Tutoring Systems (ITS). Most existing systems provide negative feedback, i.e. feedback on errors most of the time. Positive feedback that is commonly provided by human tutors in tutoring sessions is barely present in ITSs. Although some research has been done on the timing of feedback, very little advice is available on the content of feedback and the type of feedback to provide to the student. In this talk, I will present two studies done at ICTG which focused on feedback content and positive feedback. The studies show that theory-based feedback improves learning, and that positive feedback helped students learn much faster.