Anyone who still believes in the reality of race ought to spend some time reading graduate school applications. Every year my department receives a few hundred, a growing portion from students who identify themselves as of "mixed race" or fail to check anything at all, leaving me to use my sleuthing skills for clues about their ethnic heritage.

I'm not alone. In the 2000 US Census, 7 million people, 40 percent of whom were under age 18, picked more than one racial or ethnic category for themselves. Between 1991 and 2001 the number of students in higher education whose race is officially "unknown" increased 100 percent. Americans still use the language of race to identify themselves--they just don't agree about what "race" is.

Why do I spend so much effort trying to fit students into racial categories whose biological basis has been thoroughly discredited? According to literary critic Walter Benn Michaels, author of The Trouble With Diversity: How We Learned to Love Identity and Ignore Inequality, I'm engaging in a fruitless, even sinister, reactionary enterprise, one that distracts vital attention from the only social division that ultimately matters: class. How can I be certain that the "minorities" I'm admitting are truly disadvantaged, Michaels would ask, and not the children of the growing black middle class? And even if these minority candidates are economically disadvantaged, he'd continue, why do I assume that they, by virtue of their ethnicity, will bring more (or different) insights than other students?

Would that I could respond with the theoretical sophistication (though not the repetitiveness) of Michaels's book. Alas, my answer is quite ad hoc and mundane. Given the pool applying to my expensive private university, I've found that race is a fairly reliable proxy for disadvantage (at least relative to the other applicants). While I'm always on the lookout for telltale phrases like "first in my family to attend college," our application has no "p...