Assessing the impact of positive feedback in constraint-based tutors

Most existing Intelligent Tutoring Systems (ITSs) are built around cognitive learning theories, such as Ohlsson's theory of learning from performance errors and Anderson's ACT theories of skill acquisition, which focus primarily on providing negative feedback, facilitating learning by correcting errors. Research into the behavior of expert tutors suggest that experienced tutors use positive feedback quite extensively and successfully. This paper investigates positive feedback; learning by capturing and responding to correct behavior, supported by cognitive learning theories. Our aim is to develop and implement a systematic approach to delivering positive feedback in ITSs. We report on an evaluation study done in the context of SQL-Tutor, in which the control group used the original version of the system giving only negative feedback, while the experimental group received both negative and positive feedback. Results show that the experimental group students needed significantly less time to solve the same number of problems, in fewer attempts compared to those in the control group. Students in the experimental group also learn approximately the same number of concepts as students in the control group, but in much less time. This indicates that positive feedback facilitates learning and improves the effectiveness of learning in ITSs.