A college's graduation rate is such a basic consumer fact for would-be students these days that it's difficult to imagine that the federal government didn't even collect the information as recently as the early 1990s. If not for two former Olympic basketball players who made their way to Congress and wanted college athletes to know about their chances of graduating, people might still be in the dark about how well a college does in graduating the students it enrolls. Even before the Education Department settled on regulations in the mid-1990s, the definition that they were working from in the law was quickly becoming outdated. The law defined the rate as the percentage of full-time, first-time students who enrolled in the fall and completed their degree within "150 percent of normal time"--six years for students seeking a bachelor's degree. During the 1990s, higher education was in a state of rapid transition, with more students going to college part time and transferring between institutions, and more adults returning for their degrees. What were called nontraditional students then are today's traditional students. But very few of them are captured by the federal definition of the graduation rate. Every year, the method by which the government measures the graduation rate gets further and further from what's actually happening on campuses. For example, about one-third of students now transfer from the college where they started, according to a recent report from the National Student Clearinghouse Research Center. A better way of collecting graduation rates is already well known: a "unit-record" tracking system that would follow students from institution to institution for the full length of their college careers. Attempts to create such a system were defeated in Congress several years ago, in part by lobbyists for private colleges who worried about the privacy of student records. What's puzzling about the opposition by some in the higher-education establishment to collecting more accurate graduation data through a unit-record system is that under such a method the rates at most institutions would probably improve. One theory about why some private-college presidents oppose an improved system is that it would most likely raise the bottom-line numbers at public colleges. That would close the gap between the two groups and raise questions about the argument perpetuated by less-selective private colleges that they cost more but at least their students get their degrees on time.