The University ranks seventh among 105 national undergraduate programs in a revealed preference study authored by researchers at Harvard, the University of Pennsylvania and Boston University. The study ranks each college based on the number of accepted students who choose to matriculate there in relation to the other colleges to which they are accepted.
The most preferred institution in the study was Harvard University, followed by Yale and Stanford universities.
Christopher Avery of Harvard, one of the study's authors, said he and Caroline Hoxby, another of the study's authors from Harvard, had been interested in conducting such a survey for a long time. The study's other two authors, Andrew Metrick of Penn and Mark Glickman of BU, knew Hoxby and Avery from playing chess together. "Over the years the four of us in different groups (have thought) about how the college rankings are flawed," Avery said, adding they felt "it would be great to have a systematic approach to college rankings."
Because students often make decisions among a cluster of colleges such as Harvard, Yale and Princeton or a group including Cornell, Dartmouth and Brown, it can be difficult to see overarching preference trends. However, there is some overlap - "that's where you start learning about relative preference," Glickman said.
The study began in the spring of 1999 when Avery asked college counselors throughout the country to help with the survey. In advance of finding out their acceptance status, 3,240 high school students were asked to provide basic data included on the Common Application for the survey. After they were accepted, students also submitted data on the scholarships they were offered, where they were accepted and where they chose to enroll. The researchers then used the data to determine which colleges were preferred most among accepted students.
Glickman said the rankings are "surprisingly consistent with other rankings." The top 10 universities in the U.S. News and World Report ranking of national universities includes largely the same schools as the revealed preference rankings. Avery said the consistency adds validity to the preference rankings. "If we had something that completely clashed with U.S. News it wouldn't be credible," he said.
However, Brown and Amherst - which is a liberal arts college and thus not considered in U.S. News' ranking of national universities - do not make the U.S. News most recent top 10. Penn and Duke University, which did not place in the top 10 in the revealed preference rankings, are fourth and fifth in the U.S. News ranking.
Glickman said the results have provided "no clear answer" on what makes a college good, adding that the study only measured "how much the school is attractive to students."
Glickman, who has done research on chess tournaments, said he took those methods and "brought them over to college admissions."
The study has elicited both positive and negative feedback. Avery said college admissions officers have told him that it "reflects the way they think about the world." Both Glickman and Avery said they have heard complaints the survey is too much like a popularity contest. Glickman refuted this, saying the "choice a student is making is the result of an extensive thought process." Avery added he would "leave it up to others to decide if (the survey) is valuable."
One of the study's advantages, Avery believes, is its simplicity. "Nobody can disagree on whether this is the correct methodology (for what we wanted to accomplish) and what this means. ... They can disagree on whether it's useful."
While Glickman said the authors do not plan to repeat the study, he added that in "the ideal world somebody will use our method to come up with alternative rankings." Avery said an annual preference study would probably be more consistent in its findings than U.S. News. "U.S. News varies more than it should," he added.
Director of Admissions Jim Miller '73 said he thinks the study is "great," adding, "the more ... facets of an institution you can look at, the better." He said although he knew of Hoxby and Avery's work, he had not read the study until contacted by The Herald. Miller said the U.S. News formulas and factors for its rankings often change and are "subject to subjectivity," while the revealed preference study "probably has more statistical validity."
Miller said that the study was more of an academic piece and not widely circulated. Two local high school college counselors, Helen Scotte Gordon MAT '05 of Moses Brown School and Paul Tukey '66 AM '72 of Rocky Hill School said they had not heard of the study.
