Carleton University (CANADA)
About this paper:
Appears in: EDULEARN19 Proceedings
Publication year: 2019
Pages: 8872-8881
ISBN: 978-84-09-12031-4
ISSN: 2340-1117
doi: 10.21125/edulearn.2019.2204
Conference name: 11th International Conference on Education and New Learning Technologies
Dates: 1-3 July, 2019
Location: Palma, Spain
Educational Data Mining (EDM) is a research field that uses methods developed for big data analysis and improves and adapts them to educational setting to gain insight into the learning process of individual students. Understanding each student’s unique learning process helps educators provide effective digital educational content through customization. On the other hand, Learning Management Systems (LMS), widely used in most educational organizations as a vehicle to offer digital content and course management, provide a wealth of learner’s activity data by capturing learners’ digital footprints. Moodle is an open-source LMS that is widely used in various educational institutions.

The objective of this study is to analyze the data captured by LMS in order to gain an understanding of the individual learning process. In our research, we use a publicly available anonymized data set captured from a freely-offered fully-online course with 6119 students registered that has used Moodle LMS. Using data captured by the Moodle platform, we investigate students’ learning activity patterns over Attention and Participation dimensions.

Traditionally, learner’s evaluation is based on measuring participation and its outcome. Assignments submitted by students, quizzes and exams taken, discussions, and forum contributions are evaluated by instructors in assessing the learner’s course performance. However, this traditional approach may result in missing insight that can be gained by evaluating the student’s attention volume and pattern. Learner’s attention-related activities may offer opportunities for better support and can improve future performance. Examples of attention-related activities are course chapters and pages read, exercise questions answered, and group and course-related information viewed. Having more granular and detailed information on individual students' learning processes (attention and participation) can provide an opportunity for educators to offer customized support exactly where it is needed to improve future performance.

In this study, we investigate students’ activity patterns as recorded by LMS, along Attention and Participation dimensions to see if there is any correlation between students Attention activity pattern, Participation activity pattern, and grade-based performance. Our hypothesis is that LMS-based tracking of attention and participation activities can be used as an early measure of required support.

To answer the above question, we divided the student activities into Attention-related and Participation-related. We then created the attention and participation time series for each student representing those activities. We then performed a time series clustering analysis on both attention and participation dimension. Emerging clusters from time series analysis were then compared with students grade-based clusters. The result of analysis suggests that students having a high volume of attention activities in the early weeks of the course are most likely continue to be engaged throughout the course and have better performance at the end. Participation time series did not provide the same insight alone in the early weeks of the course. Considering the difficulty of detecting attention by the instructors, this suggests a unique ability of LMS-based data collection to identify weaker students and provide early support for them.
Educational Data Mining, Learning Analytics, Moodle, Assesment.