PageRanking academic labor markets

Marko Terviö has posted a fascinating new paper, Network Analysis of Three Academic Labor Markets. Here is the abstract (also available on Marko's SSRN page):

The academic labor market is analyzed as a citation network, where departments gain citations by placing their Ph.D. graduates into the faculty of other departments. The aim is to measure the distribution of influence and the possible division into clusters between academic departments in three disciplines (economics, mathematics, and comparative literature). Departmental influence is measured by a method similar to that used by Google to rank web pages. In all disciplines, the distribution of influence is significantly more skewed than the distribution of academic placements. This is due to a strong hierarchy of departments - the strongest being in economics - in which movements are seldom upwards. It is also found that, in economics, there are clusters of departments that are significantly more connected within than with each other. These clusters are consistent with anecdotal evidence about freshwater and saltwater schools of thought, although this division appears to be on the wane. There is a similar although weaker division within comparative literature, but not within mathematics.

This is a methodologically rich paper, with much to teach those of us in the legal academy who wish to evaluate our own profession by more quantitatively rigorous means. Even its incidental findings, such as the apparently greater extent of cliquish "clustering" in economics but not mathematics, as if to confirm Deirde N. McCloskey's longstanding assertion that economics is more a branch of rhetoric than of mathematics. Perhaps the paper's most useful twist is its application of Google's PageRank algorithm. (Google, of course, discloses the absolute minimum about its basic operating protocol. Ian Rogers of IPR Computing Ltd. offers a far more comprehensive and informative analysis in his paper, The Google Pagerank Algorithm and How It Works.) The upshot is that the very tools used to assess influence and connectedness on the World Wide Web can and should be applied to social networks such as the academic labor market.

I tip my hat to Dan Farber for bringing this paper to my attention.