Recasting the Feedback Debate: Benefits of Tutoring Error Detection and Correction Skills

摘要

Assessments of tutorial feedback reveal distinct benefits for immediate and delayed feedback. Immediate feedback has been shown to keep the learning process efficient. On the other hand, some research suggests that delayed feedback may foster better retention and transfer. However, unless appropriately designed, delayed feedback may also contribute to unproductive floundering and associated frustration. This paper presents a theoretical perspective which argues that feedback on the basis of a student model that includes error detection and correction skills can enable intelligent tutoring system designers to jointly realize benefits offered by both immediate and delayed feedback. Such a system would allow students to reflect on problem solving outcomes and take remedial actions if necessary. Feedback with respect to such a cognitive model can serve to guide students through error detection and correction activities and limit unproductive floundering. In this paper, we describe an experimental comparison between an intelligent novice version of a spreadsheet tutor, based on our theoretical analysis, and an expert version consistent with an immediate feedback tutor. Participants using the intelligent novice tutor outperformed participants using the expert tutor on a variety of tests: including performance on close isomorphs of training tasks, transfer tasks, retention tests, and tests of conceptual understanding.