Of all the ideas to come out of Margaret Spellings's Commission on the Future of Higher Education,  the final report  proposal that has been the most contentious inside the DC Beltway is the proposal for a unit-records database.  There are plenty of other controversial ideas floated in the commission's hearings, briefing papers, and report drafts, but the one bureaucratic detail that most vexed private colleges and student associations over the past year is the idea that the federal government would keep track of every student  enrolled in every college and university in the country. Given reports this year about the Pentagon hiring a marketing firm to collect data on teens and college students,  the possibility that Big Brother would know every student's grades and financial aid package has worried privacy advocates.
Fortunately, privacy and accountability do not need to be at odds.
The proposal for a unit-records database was floated in a 2005 report  that the U.S. Department of Education commissioned. Advocates have argued that the current system of reporting graduation data through the Integrated Postsecondary Education Data System  (IPEDS) only captures the experiences of first-time, full-time students who stay in a single college or university for their undergraduate education. How do we capture the experiences of those who transfer, or those who accumulate credits from more than one institution? Theoretically, we could trace such educational paths by tracking individuals, including your Social Security Number or another identifier to link records.
Charles Miller, who led the Spellings commission, was one of the unit-records database advocates and pushed it through the commission's deliberations. Community-college organizations liked the idea, because it would allow them to gain credit for the degrees earned by their alumni. But the National Association of Independent Colleges and Universities, the U.S. Student Association, and other organizations opposed the unit-records database, and in its current form the proposal is certainly dead on arrival as far as Congress is concerned.
There are three problems with a unit records database. The first problem is privacy. I just don't believe that the federal government would keep my children's college student records secure. An October report  by the House Committee on Government Reform documents data losses by 19 agencies, including financial aid records that the U.S. Department of Education is responsible for. Who trusts that the federal Department of Education could keep records safe?
The second problem is accuracy. I have worked with the individual-level records of Florida, which has had a student-level database in elementary and secondary education since the early 1990s. If any state could have worked the kinks out, Florida should have. But the database is not perfectly accurate. I have seen records of first graders who are in their 30s (or 40s) and records of other students whose birthdays (as recorded in the database) are in 2008 and 2010. The problem is not that the shepherds of the database system are incompetent but that the management task is overwhelming, and there are insufficient resources to maintain the database. Poorly-paid data entry clerks spend their time entering students into the rolls, entering grades, withdrawals, and dozens of other small bits of information. We probably could have a nearly perfect unit-records database system, if we are willing to spend billions of dollars on maintenance, editing, and auditing. In all likelihood, a unit-records database system for all higher education in the U.S. would push most of the costs onto colleges and universities, with insufficient resources to ensure their complete accuracy.
The third problem with such a database is that the structure and size would be unwieldy. Florida and some other states have extensive experience with unit records, and very few researchers use the data that exist in such states. The structures of the data sets are complicated, and beyond the fact that using the data taxes the resources of even the fastest computers, the expertise needed to understand and work with the structures is specialized. Such experts live in Florida's universities and produce reports because they are the experts. But few others are. There would be no huge bonanza of research that would come from a national unit-records database.
A Solution: Anonymous Diploma Registration
Most of the problems with the unit-records database proposal can be solved if we follow the advice of statistician Steven Banks (from The Bristol Observatory ) and change the fundamental orientation away from the question, Who graduated? and toward the question, How many graduated? The first question requires an invasion of privacy, expensive efforts to build and maintain a database, and a complex structure for data that few will use. But the second question -- how many graduated? -- is the one to answer for accountability purposes. It's the question that community colleges want answered for their alumni. And it does not require keeping track of enrollment, course-taking, or financial aid every semester for every student in the country.
All that we need is the post-graduation reporting of diploma recipients by institutions, with birthdates, sex, and some other information but without personal identifiers that would allow easy record linkage. Such a diploma registration system would fit with the process colleges and universities already go through in processing graduations. An anonymous diploma registration system could also identify prior institutions -- high schools where they graduated and other colleges where students earned credits that transferred and were used for graduation. Such an additional part of the system could be phased in, so that colleges and universities record the information when they evaluate transcripts of transfer students and other admissions. The recording of prior institutions would address the need of community colleges to find out where their alumni went and how many graduated with baccalaureate degrees.
Under such a system, any college or university could calculate how many students graduated and the average time to degree (as my institution in Florida already can). Any college or university could also count how many students who transferred to other institutions eventually graduated. High schools would be able to identify how many of their own graduates finished college from either in-state and out-of-state institutions. Institutions could figure out what types of programs helped students graduate, and the public would have information that is more accurate and fairer than the current IPEDS graduation statistics. All of these benefits would happen without having to identify a single student in a new database.
A short column is not the place to describe the complete structure for such a system or to address the inevitable questions. I am presenting the idea in more depth this afternoon at the Minnesota Population Center, and I have established an online tutorial  describing the idea of anonymous diploma registration in more detail. But I am convinced that the unit-records database idea is wasteful, dangerous, and unnecessary. Anonymous diploma registration is sufficient to address the most critical questions of how many graduate from institutions, and it does not threaten privacy.