Move over U.S. News and World Report. A group of economists has created a new college ratings system in which a school’s rank depends on how often students choose to attend the university over others.

U.S. News’s “Best Colleges 2005” ranks Georgetown at No. 25, but it is No. 16 on the economists’ list – between Cornell (No. 15) and Rice (No. 17).

According to Caroline Minter Hoxby, a professor at Harvard University and one of the study’s authors, Georgetown did well in the new ranking because a lot of students who had several top college options decided to go to Georgetown.

She and her three colleagues, Christopher Avery also of Harvard University, Mark Glickman of Boston University and Andrew Metrick of University of Pennsylvania, submitted their paper to the National Bureau of Economic Research earlier this month.

The study, “A Revealed Preference Ranking of Colleges and Universities,” analyzed the college choices of over 3,200 high school seniors in 2000. The researchers did not differentiate between universities and liberal arts colleges, as does U.S. News.

They instead treated each student’s options like a tournament, where a student’s final choice was the “winner.” The authors used a statistical model similar to those used to determine winners in chess or tennis tournaments.

For example, if a student gained admission to Georgetown, Brown or Amherst, and the student chose Georgetown, Georgetown would go up in the preference ranking while Brown and Amherst would go down. The more times Georgetown “beat” other schools, the higher its rank.

A more traditional university ranking, like the one U.S. News publishes annually, comes up with a set of criteria on which the school can be objectively scored. It considers factors like average SAT scores, student-professor ratios, alumni giving rates, average class sizes and acceptance and matriculation percentages. It then ranks schools by measuring them against this standard.

The study’s authors criticized some of the factors U.S. News uses to evaluate schools – namely, the matriculation and admission rates.

“These flawed measures are particularly bad because they encourage colleges to practice strategic admissions,” Hoxby said.

If an admissions department wants to increase the number of students who decide to attend its university, for example, it can simply accept more students using early decision because these students must matriculate if accepted. Similarly, sometimes, colleges refuse admission to highly or “over-qualified” students because these students are likely to have been accepted at several other good schools.

A college can also look more selective if it encourages many unqualified students to apply. If a school increases its applicant pool by a few thousand one year, and the majority of these additional applicants did not meet basic admission criteria, the school will accept a smaller percentage of students and appear more selective.

U.S. News does not take into account subjective characteristics that might make a school more appealing. The preference ranking assumes that students have done their research and know what makes a school worth attending.

“Some people might pick Georgetown because it’s in a great location,” Beth Minogue (COL ’05) said. “You can’t quantify the advantages of Washington’s internship opportunities and rich cultural life, and these factors definitely enhance a student’s education.”

The New York Times reported that Georgetown and Notre Dame may have done better in the new ranking because they both attract a “loyal Catholic following,” something that does not matter in the U.S. News ranking.

University Provost Jim O’Donnell warned against taking any ranking too seriously. He said that he’s seen people “absolutely miserable” at top five institutions, and seen people’s lives transformed for the better by “places you’ve never heard of.”

“Finding that right place is hard, and my personal advice is to use a lot of sources, not just rely on one ranking,” O’Donnell said. “I’m always glad when we score high and always unhappy when we score low, and I know these things affect people’s thinking. But we can’t be taken over by trying to `manage’ things like this.”

The complete revealed preference ranking and the rest of the authors’ findings are available online at https://www.ssrn.com/abstract=601105.