Cory Felder and Jenn Wilson
(We will be taking the second module)
Tufts Faculty Pilot Project
Introduction/background:
Tufts University’s T10 Strategic Plan, published in November 2013, states that a key to the school’s strength in a rapidly changing world will be “deepening connections across the university, and enhancing collaborations with other key institutions in the Boston area and the world who complement all that Tufts has to offer.”(1) Collaboration could certainly be defined in a number of different ways, but for a research institution like Tufts, a major indicator of connectedness is co-publication among professors. To gain a deeper understanding of the nature and frequency of faculty collaboration at Tufts University, we will examine the network formed by co-authorship within a particular department -- most likely, the engineering department.
Ideally, our analysis will serve as the first study in a series looking at connectedness and collaboration at Tufts University. A social network analysis of this type, for instance, may later be useful for answering the question “How international is the Fletcher School?”
We also hope this project reveals something about the nature of collaboration among experts more generally. In part, we chose to come to Fletcher because of the interdisciplinary nature of the program; learning more about why experts in different fields choose to work together, the costs and benefits of doing so, and the patterns of collaboration that form over time will likely be useful for our studies here and after graduation.
Primary Questions
Tufts University’s T10 Strategic Plan, published in November 2013, states that a key to the school’s strength in a rapidly changing world will be “deepening connections across the university, and enhancing collaborations with other key institutions in the Boston area and the world who complement all that Tufts has to offer.”(1) Collaboration could certainly be defined in a number of different ways, but for a research institution like Tufts, a major indicator of connectedness is co-publication among professors. To gain a deeper understanding of the nature and frequency of faculty collaboration at Tufts University, we will examine the network formed by co-authorship within a particular department -- most likely, the engineering department.
Ideally, our analysis will serve as the first study in a series looking at connectedness and collaboration at Tufts University. A social network analysis of this type, for instance, may later be useful for answering the question “How international is the Fletcher School?”
We also hope this project reveals something about the nature of collaboration among experts more generally. In part, we chose to come to Fletcher because of the interdisciplinary nature of the program; learning more about why experts in different fields choose to work together, the costs and benefits of doing so, and the patterns of collaboration that form over time will likely be useful for our studies here and after graduation.
Primary Questions
- What factors encourage or prevent collaboration between Tufts professors? Why do faculty choose to collaborate or not?
- How might patterns of collaboration in the department be characterized? Inclusive or exclusive? Frequent or rare? Routine or sporadic? Specialized or interdisciplinary?
Secondary Questions
- What methods of analysis might be useful to apply to the Tufts faculty as a whole to better understand collaboration networks?
- Are patterns of collaboration among professors in a given department reflective of collaboration at Tufts University at large? What lessons learned from examining a given department can be applied elsewhere in the university to deepen connections between faculty?
- How often do professors collaborate externally? With others in the Boston area? Abroad?
- Has collaboration between professors changed over time? In what ways is the department becoming more or less collaborative?
Hypotheses and Potential Network Measures
- We think collaboration will be more likely between faculty who share a specialty. Indeed, similar studies on co-authorship have suggested that collaboration tends to depend more on research specialty than other factors. We also think that the most influential professors -- those who have broken new ground in their fields, coined new terms, or influenced their discipline in some major way -- will tend to be the best connected, if only because others want to publish with them the most. These scholars may not have collaborated the most (degree), but they will likely have the shortest difference to others in the network (closeness) given their stature.
- Measuring centrality may therefore be useful. It probably won’t be conclusive, however; a high degree score could indicate a professor genuinely interested in collaborative research, for instance, or just that he/she agrees to be included in a publication to give it more credibility or visibility. Similarly, a non-tenured professor may push harder to work with a tenured professor than with another non-tenured professor. So degree is not necessarily meaningful as an indicator of a collaborative culture.
- Clique analysis, on the other hand, will probably be more revealing. Depending on how cliques are organized -- by specialty, background, years of experience, tenure, gender, political bias, language, professional experience, location, or even personality -- we can then begin to identify hubs of collaboration among faculty members. Professors who are members of multiple cliques may be ideal candidates for interviewing to test our hypotheses. Professors who are members of no cliques may also provide additional insight into the nature of faculty collaboration.
Data
Drawing from Tufts’ internal databases and online research (ie. Microsoft Visual Explorer) we will construct a two-mode, binary matrix that includes faculty members and publications. From there we can build a one-mode, valued matrix in which professors are connected by publications. The strength of the ties will be determined by how many times they have collaborated. We will also use attribute data for Tufts faculty that has already been collected, including time spent at Tufts, educational background, tenure or non-tenure-track, among others, and identify additional attributes that might be useful for our analysis. Depending on the results of our initial analysis, we will select several faculty members to interview and test our hypotheses.
Some potentially useful studies/sites:
http://www.pnas.org/content/101/suppl_1/5200.full
http://academic.research.microsoft.com/VisualExplorer
(1) http://provost.tufts.edu/wp-content/uploads/Tufts-Strategic-Plan-Exec-Summary-web.pdf
(1) http://provost.tufts.edu/wp-content/uploads/Tufts-Strategic-Plan-Exec-Summary-web.pdf
1 comment:
Very well done, and well-written to boot. Only thing missing is the Complication: what's made it hard for Tufts to do this previously? One big reason is that it's hard to visualize. And that sets up a Question like "Is there a way that we can get our minds around how we connect to the world so that we can take advantage of our worldwide network?" to which social network analysis is the obvious answer. Just trying to make something good even better.
You've obviously done some thinking here. Nice reference to T10 (you know who your ultimate client is,) sensible questions, and a really explicit and well-reasoned section on hypotheses and measures.
You're off to a great start.
Post a Comment