Student Modeling In An Intelligent Tutoring System.

Abstract

Intelligent tutors have often been used mainly to teach students. In the ASSISTments project, we have emphasized using the intelligent tutoring system as an assessment system that just so happens to provide instructional assistance during the test. In this chapter we review and summarize some of the main studies we have done with the system. Usually its believed that assessment get harder if students are allowed to learn during the test, as its then like try to hit a moving target. So our results are surprising, that by providing tutoring to students while they are assessed we actually improve the assessment of students’ knowledge. We also review out attempts to give teachers feedback based on fine grained skill models. Overall, we conclude that using intelligent tutoring systems to do assessment seems like a reasonable way of dealing with the dilemma that every minute spent testing students takes time away from instruction.