<$BlogRSDUrl$>

Thursday, October 21, 2004

Chess and School Rankings 

Do they have anything in common? Now they do - Four researchers (one of whom, Andrew Metrick, is a colleague at Wharton Finance:-) have used the basic method that rates chess players to rate schools. The article has been covered by New York Times and is in no need for additional publicity. Yet here's the gist of it.

Instead of measuring statistics like admissions rates or having college administrators rate their peers, the researchers analyzed the college choices of more than 3,200 high-achieving high school seniors from the class of 2000. By asking those students where they enrolled when they were accepted by several different colleges, the economists compiled a won-loss record for each college as if it was a competitor in a chess tournament. The researchers then generated a preference ranking for more than 100 colleges, employing the same scoring system used for chess masters.

The method might have some issues (e.g., using students as judges and the potential problems therein) but has one advantage that stands out. It is a much more difficult system for schools to game - a tendency that ratings such as that of US News, Financial Times etc. encourage. In Andrew's words:

"What you are getting in all these other systems is sort of an expert analysis of polling data. This provides a market-based view."