A report issued yesterday by NYU and the College Board attempts to lay out a case for why we should remove “model” from in front of “minority” with respect to AAPI students. The New York Times covers the report:

The report, by New York University, the College Board and a commission of mostly Asian-American educators and community leaders, largely avoids the debates over both affirmative action and the heavy representation of Asian-Americans at the most selective colleges.

But it pokes holes in stereotypes about Asian-Americans and Pacific Islanders, including the perception that they cluster in science, technology, engineering and math. And it points out that the term “Asian-American” is extraordinarily broad, embracing members of many ethnic groups. [Link]

I’m a census data geek so I had already flipped through a Powerpoint presentation of some of the study results before the Times write-up. I should admit that even before looking at the data I was slightly biased in that I thought it would be skewed in favor of the results that the “mostly Asian-American educators and community leaders” wanted to see. These types of studies usually seem to be. I agree that it is demeaning and not nuanced enough to label Asian Americans as model minorities, but I am also against making weak arguments to prove the contrary. Besides, if we weren’t “model minorities,” then why would Hollywood always portray us so? Here is Robert Teraishi’s presentation.

Figure by Robert T. Teranishi, N.Y.U.

Something about the above graph surprised me. It makes sense that Asian American groups (AAs only, not the PIs) with the shortest orange bars would be relatively better-off economically than those with longer orange bars. However, Chinese Americans seem to break that pattern. As a group they have a pretty long orange bar, indicating a substantial subset with poor English skills, and yet as a group they seem to have been pretty successful. I’m guessing the reason my perception is a bit skewed is that the absolute number of Chinese Americans is far greater than Indian or Filipino Americans. Maybe percentage-wise they haven’t been as successful as I assumed.

The report also said that more Asian-Americans and Pacific Islanders were enrolled in community colleges than in either public or private four-year colleges. But the idea that Asian-American “model minority” students are edging out all others is so ubiquitous that quips like “U.C.L.A. really stands for United Caucasians Lost Among Asians” or “M.I.T. means Made in Taiwan” have become common, the report said. [Link]

Ok, I got to say though, having been both a student and a T.A. at UCLA, it is kinda of true (Abhi ducks for cover). At least most students seemed to have black hair.

The next interesting plot was the one showing the double peaked Asian American students’ scores vs. the Gaussian distribution of all students. The point that the statistician is trying to make? There are both really smart and really poorly performing Asian Americans so we should assume they are all smart. My problem with this plot, and much of the other presentation, is that it conflates Asian American data with Pacific Islander data. There are many historical reasons why the two groups would perform differently. When most Americans stereotype Asians as a model minority (and poke fun at them as Hollywood characters) I’d argue that they usually have the AAs in mind and not the PIs. I mean, there aren’t many nerdy Samoan characters on TV.

Figure by Robert T. Teranishi, N.Y.U.

“Certainly there’s a lot of Asians doing well, at the top of the curve, and that’s a point of pride, but there are just as many struggling at the bottom of the curve, and we wanted to draw attention to that,” said Robert T. Teranishi, the N.Y.U. education professor who wrote the report, “Facts, Not Fiction: Setting the Record Straight.”

“Our goal,” Professor Teranishi added, “is to have people understand that the population is very diverse…” [Link]

Explaining to people the reasons for that diversity and how specifically it affects the presented data would have been useful. What would the above plot have looked like with just the AAs ploted vs. All? To be fair though, I think the author “gets it,” although I wish he sated this up front. His first conclusion is: “We need to collect and report data in a way that allows for further deconstruction of the variation within the AAPI racial category (by ethnicity, nationality, immigration history, and language background).”

There are a few other related presentations of interest that can be found on the Congressional Asian Pacific American Caucus website here.