Traditional measures of a college's student-loan default rate can omit those who drop out and defer payment. Comparing graduates to borrowers who are repaying loans provides a more accurate metric for policy makers, the author argues.