Abstract
The amount of information contained in any educational data set is fundamentally constrained by the instructional conditions under which the data are collected. In this study, we show that by redesigning the structure of traditional online courses, we can improve the ability of educational data mining to provide useful information for instructors. This new design, referred to as Online Learning Modules, blends frequent learning assessment as seen in intelligent tutoring systems into the structure of conventional online courses, allowing learning behavior data and learning outcome data to be collected from the same learning module. By applying relatively straightforward clustering analysis to data collected from a sequence of four modules, we are able to gain insight on whether students are spending enough time studying and on the effectiveness of the instructional materials, two questions most instructors ask each day.
| Original language | English |
|---|---|
| State | Published - 2018 |
| Event | 11th International Conference on Educational Data Mining, EDM 2018 - Buffalo, United States Duration: 15 Jul 2018 → 18 Jul 2018 |
Conference
| Conference | 11th International Conference on Educational Data Mining, EDM 2018 |
|---|---|
| Country/Territory | United States |
| City | Buffalo |
| Period | 15/07/18 → 18/07/18 |
Bibliographical note
Funding Information:We thank the UCF LS&T team led by Dr. Francisca Yonekura for developing the Obojobo platform for implementing OLMs and Dr. Patsy Moskal at the Center for Distributed Learning at UCF for commenting on the manuscript. The project was supported by star-up funds from the University of Central Florida.
Publisher Copyright:
© 2018 International Educational Data Mining Society. All rights reserved.
Keywords
- Clustering analysis
- Data interpretability
- Online instructional design
- Supporting teachers