Collaboration network puzzle

[David Wojick]

Dear Gohar,

Perhaps (1) there are several distinct subfields, which do not collaborate, but which are often found in the same department or institution. Either that or (2) there are several distinct schools of thought, pursuing different hypotheses, but again often found in the same department. Option (2) can be a characteristic of an emerging field that has yet to settle on a paradigm.

David

On May 22, 2015, at 2:57 AM, "Gohar F. Khan" <gohar.feroz at GMAIL.COM> wrote:

> Adminstrative info for SIGMETRICS (for example unsubscribe): http://web.utk.edu/~gwhitney/sigmetrics.html
> Dear Colleagues:
> 
> Please help me solve the following puzzle.
> 
> Using the Web of Science publication data, I have constructed two networks, 1) author level collaboration network, and 2) institutional level collaboration networks. The publications are in a particular research field.   
> 
> The problem is that the author network is composed of several small disconnected clusters with largest component consisting of only 43 nodes or authors. However, the institutional level collaboration network is well connected with 326 institutions in a single component. 
> 
> I just wonder what could explain these differences? Normally, when I construct such networks, the institutional network is sparse as compared to the authors network, as all the publications (or nodes) from a particular institute are merged into a single node when creating an institutional level collaboration network. But, here it is the opposite. 
> 
> What could cause this to happen? Any ideas? 
> 
> 
> Regards,
> 
> -- 
> Gohar Feroz Khan, PhD
> Assistant Professor
> Korea University of Technology & Education (KoreaTECH)
> ---------------------------------------------------------------------
> 
> Stay tuned for my new book on 7 Layers of social media analytics to be available soon....
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20150522/f53893ba/attachment.html>