Guest Post: Avery Haller, Davidson class of 2015
The DavidsonX MOOC research project starts with the simple question: “How, if at all, is the MOOC experiment impacting residential teaching and learning?” In response, we are undertaking a mixed method research study of two MOOC-infused residential courses during the fall semester of 2014. This case study is essentially a broad exploration intended to reveal future research questions about a Davidson course experience. The value of a small residential college lies in its ability to support deep critical learning through discourse and community; we want to see how online learning impacts those strengths.
As an anthropology major, I’m not the first person people think of when they want some data work done. Lucky for me, Davidson is a place that trusts its students’ interdisciplinary agility. This semester, I got the chance to do a social network analysis of Davidson College’s MOOC experiment. Through this project, I’ve learned how to combine quantitative and qualitative approaches to achieve more meaningful analysis. Even as an undergraduate with little previous data analysis experience, I was able to use social network analysis (SNA) to reveal a surprisingly professor-centric structure in Davidson’s online classroom.
Initially, I thought I would find that the forum was a space for students to connect to other students. However, my network analysis revealed that the Medicinal Chemistry forum mirrored a traditional classroom experience, with most of the information flowing to and from the professor. My findings, though preliminary, will be used to guide future questions about student-to-student and student-to-professor relationships in residential and online Davidson classrooms.
The goal of my research was to visualize and analyze the information flow in Davidson’s Medicinal Chemistry course using SNA in order to understand how learning happens in digital environments. I worked closely with Kristen Eshleman, Director of Instructional Technology and a DavidsonX researcher, to design the project.
SNA works well for learning analytics because it examines information flow. In the classroom this means looking at the paths along which knowledge travels. Basically, SNA posits that information moves around a network through the connections between and among people. People are called “nodes” and connections, or relationships, are called “edges.”
Methods: Making the Network
In the MOOC, information on the forum travels when users post comments, questions, and answers.
Step 1: Nodes and Edges
First, I had to identify my nodes (people) and the edges (connection between two people). I made each forum post author a node. Two authors gained an edge between them if one commented on the other’s post. Therefore, each edge represented the exchange of information that occurred when authors interacted online. The edges were weighted to reflect the number of interactions; the more often authors commented on each other’s posts, the thicker the edge between them.
Step 2: Data
Daniel Seaton, researcher and instructional designer for DavidsonNEXT, prepared the initial data files for me.
Next, I had to get it in a readable format for Gephi, my chosen network analysis software. A fellow Davidson student, Erol Cromwell, wrote a data parser that identified which comments belonged to which initial post.
Step 3: Visualizing the network
After importing my node and edge connections, the graph looked like this:
So far, not a very intuitive, or pretty, illustration.
Gephi could give me the graph’s SNA measurements right away, but it took a little maneuvering to get a helpful visualization. Gephi has a great tutorial for getting started.
In the end, Gephi and I created this picture, which convincingly visualizes the importance of the professor (the big, dark blue node at the top) in the Medicinal Chemistry forum.
Node Color: Betweenness centrality (darker = more central)
Outside circle: Betweenness centrality is above 600
Inside circle: Ordered by betweenness, counterclockwise
Node Size: Betweenness centrality (bigger=more central)
Edge Color: Weight
Edge thickness: weight
Edge direction: clockwise
Step 4: Making Sense of it All
When forming the final picture, I used the “ranking” tab to highlight each node’s “betweenness” centrality. “Betweenness” is a measure of a node’s bridging capacity. In this case, the professor emerged with the highest betweenness score by far. This means that he often connected two students who both interacted with the professor, but did not interact with each other.
Additionally, the professor had the highest number of interactions coming in, made visible by the huge number of clockwise edges leading to his node. Notably, the edges leading to the professor’s node are notably darker than most others in the graph; this means that students were more likely to interact multiple times with the professor than they were likely to re-engage with each other.
In the end, this graph greatly resembles interactions in a physical classroom. Students spend most of their time focused on the professor, posing questions and comments to the front of the room rather than directly to their peers.
But, as pretty as it is, even SNA has its limitations. Graph structure can display important revelations about classroom dynamics, but one must follow up in-person with students and teachers to understand the motivations and situations that lead to those dynamics.
In addition to gaining a deeper understanding of the value of mixed method research, I have been able to contribute a residential student voice to these interpretations. In the end, this project is just the start of an ongoing research process about how we might integrate online learning at Davidson College. I am excited to contribute data and methods to the ongoing work.
Avery Haller graduated from Davidson this May with a major in Anthropology and a minor in Religion. Outside the classroom, Avery has been a leader in Davidson’s Catholic Campus Ministry. A native of Seattle, Washington, she plans to move back to her hometown and pursue work in the nonprofit sector.
28 Mar 2019
14 Jun 2017