Detection of At-Risk Students in Programming Courses

Abstract

Since the demand for programmers is increasing, programming courses are being offered widely. In this context, students’ motivation can be damaged by difficulties they encounter in their programming courses. Although teachers’ support is necessary to prevent such an issue, it is impossible for teachers to directly monitor all students’ programming activities at the same time and determine which students have troubles with programming. Therefore, several studies have been conducted to help teachers monitor students. However, these studies do not provide an understanding of the activities of students who do not run their code, which may lead researchers to miss students who are in trouble. In this paper, we propose an indicator for detecting students who need coding support by analyzing programming logs that are recorded even when the students do not run their code. This gives teachers deeper insight into the students’ programming performance. Although further work remains, the validation of this indicator shows that it could detect those students who are in trouble.