Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Can typical behaviors identified in MOOCs be discovered in other courses?

: An, Truong-Sinh; Krauss, Christopher; Merceron, Agathe

Fulltext (PDF; )

Hu, Xiangen (Ed.):
10th International Conference on Educational Data Mining, EDM 2017. Proceedings. Online resource : Wuhan, China, June 25-28, 2017
Wuhan, 2017
International Conference on Educational Data Mining (EDM) <10, 2017, Wuhan>
Conference Paper, Electronic Publication
Fraunhofer FOKUS ()
MOOC; typical behavior; X-means clustering

The emergence of Massive Open Online Courses (MOOCs) has enabled new research to analyze typical behaviors of learners. In this paper, we investigate whether this research is generalizable to other courses that are backed by a learning management system (LMS) as MOOCs are. Building on methods developed by others, we characterize individual learning behaviors in different ways taking into account specificities of the LMS we use. We then apply clustering techniques to uncover typical behaviors in two university courses. One course, JavaFX, teaching about the software programming framework, has been offered as a supplementary online course to students enrolled in an online degree. Enrolling in this course was voluntary and students did not earn any credit towards their degree; in this sense, the JavaFX course bears similarities to a MOOC though it is neither massive nor open to everybody. The other course is a classical face-to-face course on Advanced Web Technologies (AWT) backed by our LMS; students earn a degree when they pass the final exam. It turns out that the different characterizations of individual learning behaviors are consistent for the JavaFX course and uncover typical behaviors reminiscent of those found by others in MOOCs, while they aren't as applicable to the AWT course. However, typical behaviors found in the AWT course give insights on styles that lead to better marks.