Assessing the impact of students’ activities in e-courses on learning outcomes: a data mining approach

This paper aims to study the relationship between students’ activities in the e-classroom and grades for the final exam. The study was conducted at the Faculty of Administration, University of Ljubljana among first-year undergraduate students. In the e-classroom, students learn new content for individual self-study, and their knowledge is checked with quizzes.

In the empirical study, the relationship between performance in quizzes and at the final exam was studied from two perspectives. First, successful and unsuccessful students (in terms of quizzes) were compared. Second, the Orange data mining software was used for two predictive modelling tasks. The research question was based on a student’s quiz performances, is it possible to predict whether the student will pass an exam and will the student’s grade for the exam be good.

The empirical results indicate a very strong connection between a student’s performance in quizzes and their score for the final exam in the course. Significant differences in performance were found between students who had completed most quizzes and those who had not. Moreover, the results highlighted which quizzes, in other words topics, are most important for passing an exam or obtaining a better grade. Therefore, the quality of individual study in the e-classroom positively influences a student’s performance.

The paper is the first to assess the impact of students’ activities on learning outcomes in undergraduate public administration programmes by applying a data mining approach.