Identifying and detecting abstract or latent topics that are not directly observable. Rather, the topics are inferred based on the content of a text and reveal its hidden semantic structures. As such, it produces clusters of similar words that characterize the topic.
Published in Chapter:
Two Case Studies of Online Discussion Use in Computer Science Education: Deep vs. Shallow Integration and Recommendations
Gokce Akcayir (EdTeKLA Research Group, Department of Computing Science, University of Alberta, Canada), Zhaorui Chen (EdTeKLA Research Group, Department of Computing Science, University of Alberta, Canada), Carrie Demmans Epp (EdTeKLA Research Group, Department of Computing Science, University of Alberta, Canada), Velian Pandeliev (Faculty of Information, University of Toronto, Canada), and Cosmin Munteanu (Institute of Communication, Culture, Information, and Technology, University of Toronto Mississauga, Canada)
Copyright: © 2020
|Pages: 26
DOI: 10.4018/978-1-7998-3292-8.ch017
Abstract
In this chapter, two cases that include computer science (CS) instructors' integration of an online discussion platform (Piazza) into their courses were examined. More specifically, the instructors' perspectives and role in these cases were explored to gain insight that might enable further improvements. Employing a mixed methods research design, these cases were investigated with text mining and qualitative data analysis techniques with regard to instructors' integration strategies and students' reactions to them. The results of the study showed that among these cases, one entailed a deep integration (Case 1) and the other a shallow one (Case 2). Instructors' presence and guidance through their posting behaviors had a bigger effect than the nature of the course content. Additionally, TA support in online discussions helped address the limitations of the asynchronous discussion when the TAs had the maturity to only respond to questions for which they were adequately prepared.