Specifically, Affirmative Action in college admission was abolished by 2003 decision for Grutter v. Bollinger, where Justice Sandra Day O’Connor wrote in the majority opinion that race should be considered alongside “all factors that may contribute to student body diversity.” Just like that, holistic admission with racial diversity was born.
You can call it whatever you like but it’s still illegal to have different admission standards based on race. The country has changed tremendously since 2003. Back then both President Obama and HRC said marriage was between a man and a women. The biggest romantic comedy in a decade is Crazy Rich Asians and if you go to virtually every major resort including in America the second language written on the store window is Mandarin. Times have changed and open Asian discrimination is going to be struck down.
The problem with the Asian discrimination case is going to be optics. I think everyone who has looked at, and understands, the numbers can figure out what is going on. Asians are being systematically penalized in the assignment of subjective ratings by the Harvard adcoms - in particular the Personal Rating and the Overall Rating, and then they are being subjected to more stringent standards in the final culling of applications. There is obviously a soft ceiling here, just as there is obviously a harder floor for URM admits. The result is that Asians face much lower odds of being accepted for any given set of observable characteristics (and reasonably inferred non-observable ones).
However, in terms of the admits, the stats of Asians are not that much higher than any other group in the admit pool. We have all seen things like the Espenshade study that talks about Asians needing 100-250 points higher, for instance, on the SAT. And I think that is true if we are looking to equalize odds. In order to have the same chance X% of admission as a white or URM applicant, then the Asian must have Y% or Z% higher scores, and these Y and Z are big numbers.
(Obviously, we are all using “chance” in a very sloppy, non-mathematical way. Each applicant has a fixed, but unknowable a priori chance, based primarily on those aspects of the admission process that are going to be random, but let’s not let that get in the way of the story.)
Because Harvard gets to choose from so many good candidates, we get the situation in which preferences are very large, but the observable characteristics of the actual admits are small. For instance, the “chances” of a URM being admitted with the same set of characteristics is probably somewhere (in my opinion) between +300% to +1000%, depending on which URM category and which group is being compared against. The preference will be largest as compared with Asians, and the converse of course is that the “odds” for Asians are lowest. This is the essence of the discrimination and race preference. But here is where the optics problem comes in: the actual admits will not look that different from each other on observables, because the large majority of “great candidates” in one group will be rejected while the large majority of “great candidates” in another will be accepted.
I looked closely at the SAT data for all admits to Harvard that was provided in the Arcidiacono reports. There is one clue that allows us to extract the actual score differences from the simple z-score figures provided in the appendices. In the body of the main report, the actual score differences are provided by group with respect to all applicants: Asian applicants’ average SAT scores were 24.9 points higher than whites, 153.9 points higher than Hispanics, and 217.7 points higher than blacks.
Looking at the z-score data, one observes that there is basically no difference between the average verbal score of the Asian and white applicants (using the expanded data set because the difference figures were given for the whole applicant pool). This implies that the entire 24.9 point observed difference was on SAT math, which allows the extraction of the SD of the math scores (in terms of points) for the whole applicant group. I’ll save everyone the math, but it should be intuitively obvious that once you have the SD of one of the SAT components together with the total raw score differences, then it is easy to extract the SD of the verbal section. and by extension the points differences by component score.
Bottom line, here are the differences for Asians versus other groups that I calculated (I tested these numbers in a number of ways from the data, but I would appreciate of course any additional eyes and input to either confirm or show where I went wrong):
Asian minus white
SATM – +20.5 points
SATV – +3.4 points
Total – +23.9 points
Asian minus Hispanic
SATM – +41.8 points
SATV – +23.9 points
Total – +65.7 points
Asian minus black
SATM – +55.1 points
SATV – +27.3 points
Total – +82.4 points
These SAT score differences are very small in the court of public opinion, especially as we have become used to hearing about 100-250 point gaps.
Very informative post.
But another thing that gets lost in the discussion is that the SAT has been getting progressively easier over the last 25 years. This means that there is a greater score compression at the top, and therefore there will be naturally be a lower difference between the acceptable students and the really strong students.
@hebegebe
I totally agree about the score compression at the top masking differences in ability, leading to an apparent understatement of the differences both between and within groups. The analogous argument regarding inflated HSGPA over the past 25 years is also true and has the same effects.
For the mathematically minded, there is another “artifact” of the data that should be noted. Because of the raw data (many very high scoring Asians, comparatively few high scoring URM, with whites intermediate), the distributions of observable characteristics within the admit pool are likely fairly different. Adding the adcoms’ ceiling constraints on Asians and floor constraints on URM, the Asian distribution is likely to be negative skew (left tailed) while the URM distributions are likely to be positive (right tailed). As a simple artifact of the mathematics, this combination is likely to push the means closer together than if the distributions were equivalent skew, and again has the effect of minimizing “true” differences.
But no one in the court of public opinion is going to understand that! Honestly, most of the difference between Asians and other groups is going to be explainable by “STEM focus” (just to be clear, I don’t believe this, but I do think it will resonate with the public, especially in light of the small SATV differences and the general public’s aversion to math anyway :).)
Satchel the difference right now is that the DOJ is lining up against Harvard and the other schools. This type of strong arm treatment previously was reserved for private businesses. Big business caved to this pressure long ago because the threat and cost of litigation was too risky. The same will be true here at most schools. Harvard might be rich enough to decline public money but I doubt they will choose to do this. The world has changed so much in the past 20 years. The new Chinese influence is everywhere but especially in the USA. There is no way this can continue for many reasons but the new political power of the Chinese have already been felt with the defeat of CA SCA 5. Think about it. California is a one party state but the Asian influence was strong enough to overwhelm the influence of the blacks and Hispanics.
I agree with this too. Elections have consequences, and I prefer this DOJ to the last.
Also, I am not sure that the whole “we can decline public money” is going to fly anymore. The owner of an apartment building doesn’t accept public money explicitly (leaving aside special negotiated situations like BMR or government loans) and yet she cannot discriminate on the basis of race. The traditional argument from the left is that all private business owners benefit in a general way from public monies and taxes. Wouldn’t Harvard still benefit from, say, public roads and infrastructure even if it turns down explicit Federal funding so that it can continue to discriminate by race?
Those gaps are small enough as far as standardized tests scores go (at least for Harvard) that they may have a much harder time proving systematic discrimination. There is a pretty significant gap for applicants, and if the admits had those same gaps it would be an easy case to prove. I would expect the tippy top elites to have the smallest gaps but didn’t expect the gap to be that small (although from my own experiences of knowing minorities that go to Ivy League schools, they all tested at the same levels as the admitted class).
@SAY Actually, the practice of DOJ intervening in civil right cases started in Obama era. The Harvard case was brought forth and bankrolled by Edward Blum, a Jewish American on the political right with no Asian or Chinese connections.
https://www.nytimes.com/2018/08/30/us/politics/asian-students-affirmative-action-harvard.html
Jzducol the DOJ under Obama and others have gone after private businesses for years. What’s new is using this tremendous government power against the elite colleges. With the full support of the DOJ the schools will be forced to cave.
Depending on the dynamics to emerge in the trial court, there is a nontrivial chance of a negotiated settlement of the SFFA case, with DOJ involvement and continuing supervision. The case for Asians against Harvard is strong, but not overwhelmingly so. Both sides have reasons to worry about a SCOTUS ruling, and there is still huge institutional pressure at the elites to preserve their ability to virtue signal through race balancing.
As noted by Justice Thomas in his Grutter dissent, it is not difficult to create facially neutral admissions criteria that would result in a diverse student body, perhaps even more diverse than the one it has now. The problem is that Harvard wants different sets of criteria for different races.
The most glaring hypocrisy to emerge from the documents is the treatment of disadvantaged African Americans. In both the OIR and Arcidiacono analyses, being disadvantaged was a significant negative for African Americans, while a positive or neutral for other groups. Even with full controls in both analyses, Harvard preferred the privileged African American. In plain English, that means that if Harvard had two black kids with the same adcoms ratings, grades, scores, equivalent recommendations and ECs, it would prefer the one who went to Choate and had wealthy professional parents to the one who went to the LPS and had working class parents. That should be publicized more, because it gets at the heart of what Harvard is really doing with “holistic” admissions (preserving privilege, including the elite privilege to virtue signal).
Justice Thomas’ dissent here: https://www.law.cornell.edu/supct/pdf/02-241P.ZX1
Declining federal funds to do what they want has worked for Hillsdale, Grove City and a handful of others.
About 40 more religious colleges are exempt from Title IX because complying “goes against their beliefs”.
It will be very interesting indeed if Harvard goes down that path.
“Harvard might be rich enough to decline public money”
I don’t know the validity of this, but a GC told me that if Harvard has students who get financial aid with federally backed loans or loans from the govt, they could lose that too, meaning they’d have to consider that as well.
It’s possible but Harvard is currently recieving more than 500 million in federal funding. It’s much more likely they will find a compromise with the DOJ. Either way the number of Asians students is going to rise to resemble CalTech, MIT, and the UC’s. But the endowment sizes drop rapidly and few schools could even consider dropping federal money.
Sure, but AFAIK none of those schools is discriminating on the basis of race. Harvard and the other elites are. I’m not an expert in this area of the law, but i would think there is a qualitative difference.
^ they are discriminating on the basis of other protected classes. Laws regarding sexual orientation and gender seem to be the main concern for them.
They have a LOT less money than Harvard to work with, too.
Grove City, Hillsdale, etc also decline Pell and federally backed loans.
NYT SCOTUS reporter discusses the coming end of AA
It will get interesting when POTUS nominates a conservative female person of color for his next pick.
If he does, it will be interesting to see if people critique/accuse her for having personally benefited from AA??? Can of worms?