It’s been an ongoing joke between fellow gnovis blogger, Trish, and I that a semester full of stats and social network analysis has seduced me into post-positivism. It’s true. I’ve learned that I like to measure things.
But in all seriousness, quantitative naysayers out there should consider the benefits of visualizing data. On the one hand, charts, graphs, and indexes are limited by their simplicity, and can often hide nuances and complexities. On the other hand, these same tools can be quite powerful for illuminating patterns previously indiscernable among data sets.
Take the CCT cirriculum. In the Fall 2009 semester we have classes offered in seven clusters: Cultural Studies, Issues in Globalization, Media, Art & Representation, Media and Politics, Technology & Society, Technology, Business & the Economy, Technology & Information Policy. What can we learn about the structure of our clusters, and the relationship between CCT clusters and courses? As a midterm project for social network analysis, we were asked to do just this. And what I learned through my analysis may (or may not) surprise you.
Behold. The CCT cirriculum as a network graph:
This graph was generated in NodeXL, an addon for Excel (if someone knows of a mac friendly sna software tool, please share). The black, square vertices represent each CCT course. The sphere vertices represent CCT clusters (color-coded for ease of reading). Each line, or edge, represents a connection between a course and a cluster, i.e. anytime there is a link between a black vertice (course) and a colored hub (cluster) we know that course was included in that cluster.
As we can see, Issues in Globalization is the most supported cluster with a degree of 13 (degree refers to the number of links that exist between a single vertex and other vertices). Cultural Studies is the second most supported cluster with a degree of 12. The least supported cluster is Media, Art, and Representation with a degree of 5; Media and Politics is a close second with a degree of 6.
One could argue that the Technology and Society cluster has the highest betweenness centrality, in that it has a high degree and is linked to other highly connected courses (such as What’s Shaping the Wired World, Intellectual Property, Looking at Photography, Communication and the Public Sphere, and Communities of Practice). This means that while Technology & Society may offer less courses than other clusters, it is more central. Technology & Society courses could act as bridges between many other subject areas.
We could also speculate that Media, Arts and Representation has a low degree (i.e. offers fewer courses) because many of the disciplinary influences are similar to Cultural Studies. Media and Politics, on the other hand, is more influenced by the discipline of political science than other clusters. In short, people who complain that Media & Politics gets the shaft…are in a way vindicated. But this may be less about institutional support and more about a lack of disciplinary overlap. That’s simply speculation on my part….Do readers have alternative thoughts/explanations?
While this project may seem basic, consider the dramatic difference between the information represented here:
Same data, different vatage point. So, no matter your taste for objective quantification, you have to admit: post-positivism can be really pretty…