June 9, 2013

Let's Not Have More Disaggregated Data

Quite a few people have built careers in higher education around the supposed need to study how different groups compare, and when the inevitable disparities are discovered, setting up programs to address the "underrepresentation problem." To get a sense of just how deeply ingrained such thinking is, consider this piece from Inside Higher Ed, "The Deceptive Data on Asians."

In it, we learn that a recent study by ETS and a group called the National Commission on Asian-American and Pacific Islander Research in Education has demanded that colleges and universities collect and report disaggregated data about Asian-American students "as much as possible."

We need such data because Asians have been cast as "the model minority" and therefore beyond the purview of all our "affirmative action" policies. Once you disaggregate the data, however, you can find all kinds of imbalances and inequities that cry out for attention.

If you look at the charts in the story, you see that there are huge differences in educational attainment between students with different Asian ancestries. For example, those with Hmong ancestry have much lower educational levels (only 14.7 percent having earned a B.A. or higher) than do those of Taiwanese descent (74.1 percent). Now we can see that there are serious problems that have been, in the words of Professor Robert Teranishi of New York University, "overlooked and misunderstood."

Obviously, we need more "outreach" to the groups that are "underrepresented."

For the sake of argument, let's take this idea seriously. The disaggregation proposed doesn't go nearly far enough. Colleges and supposed to report, e.g., "Sri Lankan" as a category, but Sri Lanka is a badly divided country with considerable inequality among its five main ethnic groups: the Sinhalese, Tamils, Muslims, Burghers, and Veddah. It's likely that there are imbalances lurking in the data for Sinhalese Sri Lankan-American students and Tamil Sri Lankan-American students. Don't we need to find out?

And don't forget the possibility of sub-dividing those groups to find still more inequalities.

After reading that report, college administrators are no doubt envisioning the prospect of creating new programs and offices to run them. UC-Berkeley's Division of Equity and Inclusion might, for example, expand to address the needs of students of Laotian, Bangladeshi, and Filipino heritage, and maybe even the more "disadvantaged" groups of Taiwanese once they've been identified. The disaggregated data could be the investment capital for a new growth industry.

Instead, I suggest that we welcome the study as an occasion to reflect on the folly of grouping people according to race, ethnicity, social class, religion, or anything else, and then assuming that any group differences indicate problems we must solve. Relatively few Americans of Hmong ancestry have earned college degrees, but there are no official barriers to prevent more from doing so. The existence of the array of educational opportunities in California and other states is known to those people and if most don't think that more education is best for them, that's fine. If and when more of them want to attend college, they'll do so.

Statistics are often used as the excuse for government meddling. That's true for unemployment, trade, housing, and emphatically so with regard to education. Putting students into smaller and smaller pigeon holes on the basis of their background is unnecessary and divisive. Time to stop doing so.

Post a comment

Leave comments here. Unless they are vicious or obscene, they will be printed.


Published by the Manhattan Institute
The Manhattan Insitute's Center for the American University.