Using Big Data to Rank Universities by Field

The increase in the cost of college and university and the difficult job market have increased the demand for college rankings. College Scorecard, the U.S. Department of Education’s entry, includes information on tuition, graduate rate, loan default rate and by 2015 it is scheduled to have information on graduate earnings. The Washington Monthly has a best bang for the buck ranking which works similarly.

A new and interesting entry into this field comes from LinkedIn which uses data on its 300 million members to define desirable employers and then rank universities based on getting their graduates jobs with those employers. The methodology is somewhat opaque and a bit sketchy but the idea is to define desirable employers by industry based on the revealed preference of employees in LinkedIn. In particular, firm A is raised relative to firm B if more people move from B to A than from A to B and similarly if firm A retains its employees longer than firm B. The percentage of a college’s recent graduates who obtain employment from the desirable employers is then used to rank the universities. No cost factors are included.

The results are not too surprising although by this ranking Georgetown university does better in finance (coming 3rd after the University of Pennyslvania and Yale) than I would have expected.

The results are less important, however, than the idea of using lots of data, often collected for other reasons, to unlock hidden value. Facebook has considerably more profiles than LinkedIn, often with education and employment data, so this type of analysis could become more common and more precise.

Hat tip: Tom Acox.


Comments for this post are closed