When colleges look to compare themselves with others, they're not much different from high-school students chasing popularity: Everyone wants to be friends with the Ivy League, but the Ivy League is really picky about whom it hangs out with.
Each year colleges submit "comparison groups" to the U.S. Department of Education to get feedback on how their institution stacks up in terms of finances, enrollment, and other measures tabulated in the Integrated Postsecondary Education Data System. The groups sometimes represent a college's actual peers but more often reveal their aspirations.
Use an exclusive Chronicle tool to see who colleges cite when they submit "comparison groups" to the U.S. Education Department.
The Chronicle analyzed the relationships of nearly 1,600 four-year colleges that make up those groups to map out the power players in higher education.
The typical college selected a comparison group of 16 colleges with a higher average SAT score and graduation rate than its own, lower acceptance rate, and larger endowment, budget, and enrollment.
The eight Ivy League colleges among them chose only 12 institutions outside their own number as peers—not surprisingly, often including the University of Chicago, the Massachusetts Institute of Technology, and Stanford University.
But 55 colleges outside of the Ivy League selected at least one member of that group for comparison. Some of those colleges, such as Tufts University and New York University, are rich private research institutions.
Other colleges selecting an Ivy League institution as part of its comparison group were bigger surprises: Alabama A&M University, which selected Dartmouth College, and Regent University, which selected Harvard University, Princeton University, and Yale University.
The University of Phoenix's Jersey City campus selected 74 four-year colleges as peers, including six Ivy League institutions. (It left out Brown University and the University of Pennsylvania.)
Institutional-research officers acknowledge that their institutions' comparison groups often include desired peers that are not true peers. Colleges want to receive data reports on enrollments, graduation rates, student costs, faculty, and budgets for institutions they aspire to be more like.
The University of Delaware included not only seven other public research universities in its comparison set of 12, but also ambitiously added a few elite private universities, such as Brown.
Heather Ann Kelly, director of the Office of Institutional Research at Delaware, says the university conducted analyses, based on size, finances, graduation rates, and other variables, to come up with peers and aspirational comparison institutions.
The group of 12 colleges that Delaware submitted to the Department of Education is a mix of these, she says, and the university makes a point of referring to them as "comparison institutions" rather than "peers."
When it assesses data from comparison institutions regarding finances, research, admissions, and other measures, Ms. Kelly says, the university wants to look at colleges it wants to be more like.
"If you took a look at your actual peers, the likelihood is that you stand up pretty well with them," Ms. Kelly says. "In order to make progress, you want to be shedding light on not just your strengths, but also your weaknesses."
Members of the Ivy League are not the only sought-after comparison institutions. The most commonly selected peer institution was one of the most selective liberal-arts colleges in the country: Carleton College, which was chosen 61 times. Tied for second were Oberlin College and Davidson College, with 56 selections each.
The 107 most intensive research universities, as classified by the Carnegie Foundation for the Advancement of Teaching, also tended to choose one another as peers. Among them they selected only 65 institutions outside their number as peers, while 234 other colleges chose one of those intensive-research institutions.
Bucknell University, a baccalaureate college, selected a diverse group of research universities, master's-level institutions, and liberal-arts colleges for its comparison group of 13 institutions.
Jerome Rackoff, a former assistant vice president for planning and institutional research, says Bucknell established factors that distinguish it—such as being a large baccalaureate college and having Division I sports—and selected colleges that shared those traits.
Some of them, he admits, are "aspirant schools" because they have traits that Bucknell wants to emulate, or because they are more selective. Mr. Rackoff says Bucknell uses the comparison group to measure how it stacks up on key metrics, such as faculty salaries and student costs.
"There are a few aspirational institutions for us, like Dartmouth, but there are a number of institutions on the list that are similar to us on a number of critical metrics," he says. "The University of Richmond is very similar to us in its size and curricular structure. Colgate is similar to us in terms of its institutional type."
The peer groups submitted to the U.S. Education Department are not the only ones colleges collect. They use other groups to benchmark presidential salaries, for example. But institutional-research officers say the former groupings are often more generic.
Randy Swing, executive director of the Association for Institutional Research, understands that colleges choose aspirational peer groups because they want to match what other good colleges are doing. But reaching too high can make any data analysis "muddy," he warns.
"You could have an aspirational group that includes Harvard," he says, but "the truth is you learn more if you benchmark yourself against closer peers than that."
While institutional-research officers at Delaware and Bucknell systematically chose their peers, not all colleges put forth much effort, adds Mr. Swing. Many institutional-research offices, he notes, are staffed by part-time faculty who are not trained in making sophisticated decisions about benchmarking and comparison groups. In some cases, college leaders may cobble together a peer list driven in part by the college rankings in U.S. News & World Report.
Mr. Swing emphasizes the importance of appearances: College should realize the extent to which, at face value, a group seems like a reasonable choice.
"This is a real reminder to institutional leaders," he says, "that the data that their institutional-research office is reporting to Ipeds is publicly used and becomes the public face of their institution."