Getting College Admissions Wrong Yet Again

In a Washington Monthly essay, The College Board Capitulates to Trump, Richard D. Kahlenberg ignores what to my mind are the two most convincing discussions of selective college admissions programs since their start a century ago, Alan Dershowitz's 1991 Chutzpah and Jerome Karabel's 2005 The Chosen. Both argue that as Jews became more numerous, prosperous, and influential in the US after the late 19th century, this created a problem for the Ivy League, which had been a bastion of upper-crust WASP society since before the Revolution.

More and more Jews had the money and cultivation to send their sons to Harvard, Yale, and Princeton, with the result that their student bodies were becoming noticeably Jewish. Selective admissions programs simply began as way to limit this trend. An obvious way to do this would have been to require photos with applications and reject applicants with aquiline features, or simply to cull those with identifiably Jewish surnames, but that would be too obvious, and it might not even be fully effective.

Instead, the Ivies and other elite schools developed techniques that have persisted to this day. One of the most pervasive has been to promote "geographic diversity", to advantage applicants from the Rocky Mountain and Plains states over applicants from Northeastern cities and suburbs, where many more would likely be Jewish. At the same time, applicants from legacy families, the traditional WASP upper crust, as well as the offspring of major donors, politicians, and celebrities, were also placed in separate, advantaged categories.

Jerome Karabel went so far as to examine the specific admissions office method that puts these policies into practice: each application, let's say from a hopeful in Fargo, ND as an example, is placed bn a separate "basket", where he or she competes only with other applicants from the Rocky Mountain and Plains states. On the other hand, an applicant from Great Neck, NY, with the largest Jewish population on Long Island, will go into a "basket" of other applicants from prosperous Northeastern suburbs, many of whom can be assumed to be Jewish. Applicants from Boise, Cheyenne, or Rapid City will never compete with applicants from White Plains; they will all go up a different silo.

For that matter, applicants who happen to come from the Hamptons, Bedminster, or Potomac will also go into a different silo, but this will less likely be based exclusively on location than their family surname and how much the family has donated to the institution. But all the same, they will never compete with applicants from either Great Neck or Rapid City, and their chances of admission are considerably higher no matter what.

I've looked at the college admissions rat race here from time to time, and what I've consistently found has been that the only competitive and merit-based component of the process has been the particular silo of applicants from the Northeastern cities and suburbs, with numerous other silos deliberately designed to supply an applicant pool with limited numbers of Jews. As I said this past May,

If you go looking for the percentages of each such set-aside in typical Ivy entering classes, you'll get something like 80% who've come in via set-asides. This has me questioning how Princeton comes up with an "overall" admissions rate of 4.62% if as many as 80% of its students come in via set-asides where the rate is much higher.

But let's get to Mr Kahlenberg's argument:

A Trump Department of Justice memorandum, for instance, has declared that “criteria like socioeconomic status, first-generation status, or geographic diversity must not be used” if a university’s goal is to further racial integration on campus.

Given the president’s appalling history on matters of race, this development, while troubling, is not particularly surprising. What is mystifying is that a pillar of the higher education establishment recently went along with Trump. Earlier this month, the College Board, which administers the SAT, announced it would stop making a tool called Landscape available to colleges, which is designed to help identify high-achieving low-income students of all races.

. . . Landscape, as the College Board noted, “was intentionally developed without the use or consideration of data on race or ethnicity.” Instead, it allowed colleges to consider a student’s achievement in light of the socioeconomic makeup of his or her neighborhood and high school. Neighborhood factors included median family income, typical educational attainment, the share of families headed by a single parent, and crime rates. High school factors included the share of students eligible for subsidized lunch, the proportion taking AP exams, and the average SAT score. The idea was that if a student does pretty well academically despite these educational challenges, they have something special to offer.

Well, early in my tech career, I worked for the Los Angeles City statistical office. The "neigborhood factors" Mr Kahlenberg cites, including "typical educational attainment, the share of families headed by a single parent, and crime rates", simply correlate closely with African-American parts of town. All this is doing is identifying African-Americans by a different set of criteria, and it looks like both the Trump administration and the College Board have gotten wise to this particular con -- in effect, it puts African-Americans in a separate set-aside, has them compete with each other, and rewards this particular set of winners irrespective of how they measure up against anyone else. Mr Kahlenberg tries to answer this problem:

As Raj Chetty and his colleagues have found, living in a neighborhood with a large share of single-parent households predicts opportunity in America. A student of any race who lives in such a neighborhood and nevertheless does fairly well shows grit and determination. The fact that, on average, Black students face this extra disadvantage is hardly a reason to ignore this factor.

But here he's simply admitting that he wants either to create or maintain a set-aside for students "of any race" who live in such communities and "do fairly well". But there's another silo for students who in fact live in prosperous suburbs and wind up in top percentiles for standardized tests and other measures of achievement. As I've noted, the elite universities, with their elaborate systems of admissions set-asides, wind up accepting this demographic in only 20%-30% of any entering class. Mr Kahlenberg is simply arguing for yet another set-aside that will keep the high-achievers in the minority in even the most selective schools.

What would an elite-school student body look like if an admissions office simply dropped all pretense and admitted only the top performers, rather than admitting as much as 70%-80% via set-asides? What would the student experience be like?