Publication Date
2022
Document Type
Thesis
Committee Members
Adrienne Traxler, Ph.D. (Advisor); Ivan Medvedev, Ph.D. (Committee Member); Jason Deibel, Ph.D. (Committee Member)
Degree Name
Master of Science (MS)
Abstract
There is much research about how actors and events in social networks affect each other. In this research, three network models were created for discussion forums in three semesters of undergraduate general physics courses. This study seeks to understand what social network measures are most telling of a online forum classroom dynamic. That is, I wanted to understand more about things like what students are most central to the networks and whether this is consistent across different network models. I also wanted to better understand how students may or may not group together. What relationships (student to student, student to instructor, etc.) are formed, centralization, various clustering and correlation coefficients, and how participation in a forum unfolds were all things that were examined in this data set. Network model construction and measuring how these constructions may affect student interactions was another focus of this study. These attributes are analyzed among individual semesters, but also compared/contrasted across all three, to see if they maintain across different network models. It was found that in general as models increase in connectivity, a rise in network measures like centralization and average degree was observed. A drop in network measure values such as average vertex-vertex distance and diameter was also seen. Finally, it was discovered that changing a model from undirected to directed made an appreciable change in average degree outcomes. Overall, this research gave an appreciation of different network model construction and how different network measures may help describe social networks. It was discovered that centralization metrics may be more telling of social networks than what was anticipated. Average degree, average vertex to vertex distance and diameter followed trends we would expect to see. Other measures looked into were transitivity, average Barrat coefficient and degree correlation coefficient.
Page Count
53
Department or Program
Department of Physics
Year Degree Awarded
2022
Copyright
Copyright 2022, all rights reserved. My ETD will be available under the "Fair Use" terms of copyright law.
