Washington — Researchers use federal data going back to the early 1990s to track how student borrowing has changed over time, but adjustments made to the most recent federal study of student aid render such comparisons inaccurate — for now.

For years, the National Postsecondary Student Aid Study calculated student debt based on loan commitments, or essentially how much money students had been promised, rather than disbursements, or how much they actually borrowed. But with the 2007-8 study, the National Center for Education Statistics, the Education Department’s statistical arm, decided to calculate the loan data based on how much students actually took out, according to a notice posted on its Web site. The change made it impossible to compare the most recent data to past studies.

The statistics center is recalculating the older studies using the new method, but only the 2003-4 study has been updated so far. The National Postsecondary Student Aid Study is the gold standard for data on student aid and debt, so a wide range of research is likely to be affected. For example, a recent Education Sector report on the growth of student loans from the early 1990s to 2008 used the incompatible data and will very likely have to be redone.

The changes are significant enough that older results should be recalculated, said Mark Kantrowitz, publisher of the FinAid Web site. For example, according to Mr. Kantrowitz’s calculations, the old 2003-4 data showed that 65.7 percent of graduating seniors at four-year institutions had student-loan debt and that, on average, they had loans totaling $19,237. With the new 2003-4 data, those numbers dropped to 64.5 percent and $18,630. —Elyse Ashburn