One School, One ID

The Chronicle of Higher Education’s default rate package this week has a chart showing 20 for-profit colleges that would be at risk for losing federal aid eligibility if a new calculation were in place today. It’s a fine chart, but it also encapsulates a huge weakness in the way cohort default rates are calculated.

The U.S. Department of Education uses two main ways to identify colleges and universities–a six digit identification number that is used for the IPEDS data system and an eight digit number used by the Office of Postsecondary Education known as an OPEID. Each number is used to obtain different kinds of data. IPEDS includes information on enrollment, graduation rates, finances, and most everything else that shows up on college navigator. The OPEID meanwhile, is used for all the borrowing data and default rates on the Federal Student Aid Data Center.

Unfortunately, the OPEID and IPEDS numbers are different, so any attempt to analyze borrowing data and other school information requires matching the two figures. That’s doable, but here’s where the fun begins—a single OPEID can represent multiple schools in IPEDS.

Take for example Westwood College-DuPage, which is located in Woodbrige, Ill. The Chronicle picked out that school because of its high default rate. But in reality that rate actually represents two other colleges as well–Westwood College Forth Worth and Westwood College Atlanta Midtown–none of which is remotely close to the others. It’s impossible to judge if students from the DuPage branch are defaulting at high rates because there’s no way to separate them out. Or maybe the rates are low for those at DuPage but really high in Forth Worth. Who can say?

Sometimes the schools that are caught by a single OPEID aren’t even under the same name. WyoTech at Long Beach is also on the Chronicle’s list, but that also includes Everest College-City of Industry and Everest College-West Los Angeles. (The Kaplan schools are the worst in terms of this combination.)

But it gets worse. The University of Phoenix has roughly 73 campuses plus its online division yet reports all its borrowing in a single line. Not only does this wreak havoc with Arizona’s date when looking at borrowing figures by state, but it means the school’s 9.3 percent default rate is of minimal value. Presumably some Phoenix schools have much higher default rates and others have much lower ones. That would be great to know and given the school’s uniform curriculum would make for some fascinating research opportunities, but who really knows who does well and who doesn’t? (To be fair, the entire Penn State system does the same thing and it’s equally annoying.)

It’s easy to dismiss this issue as just a random data complaint, but it matters. Default rates are one of the few meaningful pieces of information available to consumers, aggregating them in a seemingly arbitrary manner across geographically diverse campuses undermines their usefulness. And this setup makes it impossible to factor in borrowing effectively when looking at things like graduation rate and enrollment makeup, making accountability around for-profit institutions even more opaque.  And maybe aggregating all these campuses together hides some high default rate campuses that should be kicked out of the program.

Student loans and grants are a major component of federal involvement in higher education. As such, we should have the ability to judge the quality of that investment by looking campus by campus. Arbitrary geographic aggregation or even just combining all schools to a single line is an impediment to that and should be stopped.