Combining data from a LMS and a student register for exploring factors affecting study duration

Abstract

Institutions of higher education possess large amounts of learning-related data in their student registers and learning management systems (LMS). This data can be mined to gain insights into study paths, study styles and possible bottlenecks on the study paths. In this study, we focused on creating a predictive model for study completion time estimation. Additionally, we explored the data to find out what features may affect the rapid completion of studies for a bachelor’s degree in an institution of higher education. We combined data from two sources: the Moodle LMS and a student register. The study exploited data from the entire study duration of the students. The data we extracted from the Moodle LMS focuses on the student’s diligence in respecting assignment deadlines. Based on the data, we created a model for predicting study duration and achieved an accuracy of 72%. According to this study, among the factors that may be influenced by the student herself, we found out that the most important predictors for fast study completion are a study pace that is more intensive at the end of studies, submitting assignments well before deadline and having a considerable amount of the grade 4