"Race" in College Applications FAQ & Discussion 12

Applying to College
2041

@Data10 - In your post#2040 it seems to me that you are misquoting the actual document that @SatchelSF referred to and effectively mischaracterizing the process. It is very clear, on page 140, that the three “Standard Strong” black candidates were the ONLY Standard Strong black applicants in a pretty large sample-- that included thousands of applications (10% of total applications). Harvard only wanted to look at 10% - I am sure the plaintiffs would have been happy to have the court order them to look at every one of the application sheets.

The process could not be clearer, and here is the actual explanation from page 140 of the source linked by @SatchelSF:

(quote) Harvard readers use the label “Standard Strong” to characterize an application that had strong qualities but not strong enough to merit admission. Harvard was ordered to randomly select 10% of the domestic summary sheets of applicants for the Class of 2018[meaning thousands]; to search those summary sheets for particular keywords, including the phrase “Standard Strong”; and to produce (to SFFA) the summary sheets that included those terms. [footnote omitted] Harvard ultimately produced 256 summary sheets that included the phrase “Standard Strong” for domestic applicants who were either white, African American, Hispanic, or Asian American….Table C.1 shows the rate of being labeled “Standard Strong” by race/ethnicity for domestic applicants as well as the characteristics of applicants labeled “Standard Strong”.

@Data10: you can see that your quote is not accurate, and the main ideas behind it actually come from the next section of the report, 3.2 of Appendix C, which deals with a qualitative examination of 480 different files (complete, not just the summary sheet) and is largely redacted. Read Section 3.1 (p.140) quoted directly above for the information about how the “Standard Strong” candidates were identified. It really does look like there are only three black candidates out of THOUSANDS of sheets examined (10% of the total), which is really not that surprising since Standard Strong is not used for admits but rather for those who are “close.” The average (1240 total SAT) is actually quite low, I agree. Maybe one of the black candidates had a 1000 and the others 1300s? That we can’t know.

Also, while the z-score data you posted are mildly interesting, the real question is how black admits compare to white and Asian admits, not really to the overall pool of applicants, the great majority of whom are rejected. Someone who is better at statistics than me should extract the comparative data, both for the baseline applicants (who do not include preference legacy, development and athletic admits) as well as for the expanded pool of applicants (which does include those groups). It should be easy to compare at what percentile of the white and Asian admits the average black admit falls on each of those measures (comparisons on another measure, left out of your post, the Academic Index, are very illuminating. All the z-scores are in the Appendix under B.3.1 and B.3.2 in the documents (the rebuttal reports contain some slightly revised tables as well). With that info out there now, people will be able to calculate themselves where the black admits fall in comparison to white and Asian admits. I hope this makes sense.

2042

Fair enough. I haven’t made it to page 140 of that document yet. I was quoting early pages that did not include the additional context and originally read the “cherry-picked” comment in a different document – https://int.nyt.com/data/documenthelper/43-sffa-memo-for-summary-judgement/1a7a4880cb6a662b3b51/optimized/full.pdf#page=1 .

That said, taking the average score of 3 students is still not a representative sample. It’s also not a meaningful measure of Harvard admissions criteria for Black students, as the 3 “standard strong” students are applicants that according to the document linked above “* that had strong qualities but not strong enough to merit admission.*” Maybe those “strong” qualities are some of the many admissions criteria besides SAT scores, and low SAT scores were one of the key factors that led to their rejection.

What we do know is that the many Black students who were admitted to Harvard had very different average scores from those 3 students, who were presumably rejected. It would be statistically impossible to have an minimum admission score threshold near the scores of those 3 presumably rejected students while still maintaining the this average score and the low acceptance rate…

The earlier data I posted touched on this information. Among the full group of admitted students, the average Black and White SAT section scores in 2017 were ~720 and ~745 respectively. The other numbers I listed were from the baseline sample with athletes, legacy, children of faculty, and dean/director’s list removed. Among this unhooked group, you can estimate the relative differences by looking at the listed standard errors I listed in parenthesis. For example, the linked numbers suggest in a normal distribution, 29% of unhooked (except URM) Black admits would have a higher SAT CR than the average unhooked White admit… a difference in scores that is reasonably consistent with the average section scores listed above.

2043

Back to a more generic conversation for a minute.

Does anyone think that U of Chicago, in part, eliminated the SAT/ACT score requirement so that there is one less mandatory “objective” measure, therefore giving them more freedom to accept the students they want to? I’m not criticizing it either way.

I doubt we will ever see Harvard go Test Optional (but I guess you really never know) but if they were test optional, even though most applicants would have submitted scores anyway because so many of them have high ones, would it strengthen their argument in this lawsuit or future ones because they could say the score (possibly the MOST if not only objective measure out there) was an optional component?

This thread isn’t only about Harvard and as more and more top schools become test optional, I think this is going to be a hot topic, especially as it relates to race and other socioeconomic factors.

2044

UChicago plays all sorts of games to improve its rankings and this is just its latest move. If it continues to move up its ranking, some other elite colleges (including Harvard and Yale) will likely follow, but not STEM-focused colleges. Caltech definitely won’t follow. MIT and Stanford are also unlikely to follow.

2045

Yes, I agree U of Chicago plays games with admissions more than most, but I do think it gives them more “defendable” discretion when it comes to the holistic review process.

I also think most applicants will submit scores because if they don’t they would be worry it would look like they aren’t even at the 25% mark, and they would worry about the implications of that. A tricky thing for any applicant who is applying to any school that has a TO policy.

2046

Highly qualified applicants without hooks will continue to submit their test scores, but less qualified applicants with lower test scores (including many with hooks) likely won’t. The effect of this change will certainly increase the number of applicants, making UChicago appear to be more selective (larger denominator). Another benefit to UChicago is to make the test scores of its enrollees appear higher than they would be otherwise, which also improves its ranking.

2047

completely agree @1NJParent I expressed similar opinions on U Chicago’s admissions games a few pages ago. They aren’t the only ones, but they are among the more bold with their attempts to maintain their #3 Rank in US News.

With that said, I do think there is some altruism in there and it’s not solely to maintain their rank. But that’s a big part of it…keeping up with the Ivies and the like.

2048

@data10 That there were only 3 students does not make their average not “meaningful” Those were the only 3 of 1000s randomly reviewed. It would be like saying we are going to pick the three highest kids out of 1000s, and then disregard their average because there are only 3 of them! It’s not the number of data points that invalidate a sample, it’s the SIZE of the pool from which they were drawn that gives one confidence in the validity of the sample. Here it is 1000s (10% of the applicants in a given year). Note that there are only 256 standard strong in that whole pool who are black, asian, white or hispanic (There are probably a few others who are race not listed, mixed, or native american, etc., so the true number is probably maybe 300 or so).

This “standard strong group” is clearly the wait list kids (or there is substantial overlap). That black students averaging 1240 on the SAT were included in this group suggests 1) the talent in the black group is very thin, and 2) the black kids with much higher scores (say, 1350s or up) are very often offered admission so long as their other aspects of their applications are good. Standard strong is not just a measure of SAT but of the WHOLE application - SAT is just one part… However, the fact that admissions are willing to assign standard strong to such weak score candidates is telling, that’s all.

Harvard is looking for a reason to admit black kids. A lower score will be excused in favor of a higher GPA, and vice versa to some degree. Black achievement test scores are very low (you can see that from the z-scores) and so are their math scores, on average (ditto). On each possible component of the academic index, they are relatively close to whites (relatively close for blacks, that is), but there must be a lot of variation. That is why, for instance, the components do not look so bad individually (for instance, only 64% of white kids and 67% of Asian kids had higher GPA than the average (see the spreadsheet on the “Baseline” sheet) but when you put them all together all the individual poor scores drag down the academic index of black kids to the point that 87% of whites are above their average and 97% of Asians. Since we can see that the GPA is not that far apart (and we know this from rampant grade inflation and blacks being given easier grades anyway at all these schools), why is the academic index so weak? The reason is: SAT subject test scores, horrible math scores on the SAT, AND the fact that the Asian students, and to a lesser extent the White students are being held to an insanely hard standard that every component of their academic index must be very strong. In other words, while Harvard is looking for a reason to ACCEPT decently strong black applicants, they are looking for any reason to REJECT strong whites and Asians.

This is obvious by looking at admit rates by academic index decile. Look, it’s hard for any student to be admitted to Harvard–or Yale or Princeton for that matter, we know that. Even for black kids, scores are not enough - they need to have something else. BUT, you can see that Harvard goes out of its way to “find” that something else for black kids with even decent academic indexes. Look at page 47 in that document. The admit rates for blacks in the top deciles is like 5-7 times the rate for whites and Asians! That table is one of the most telling in the data set. (One could argue that it is not just Asians being discriminated against) Now, look two pages earlier on page 45. There are only about 260 black kids in the WHOLE dataset that are in the top two deciles. Remember academic index is SAT+GPA+SAT2 test scores (and so you need to be very strong on all THREE to be at the very op–i.e a 240) ). That is 6 years work of data so we are talking about only 40-45 black kids each year! Now, stay on page 45 and notice something curious about the numbers of applicants by decile? Look at the kids who have no chance - decile1, meaning probably SAT around 1000-1100, weakish GPA, very weak achievement test scores, or some combination thereof. Although there are more than 5 times as many white applicants as blacks, there are more than twice as many blacks in that low decile! Now look at Asians - notice how the number rises through the deciles? This is all self-selection issues. Whites (one infers) understand how hard it is to be admitted to Harvard so poor scorers and weak academic kids generally do not apply in huge numbers . Asian know it is impossible for them (unhooked) so the strongest apply in greatest numbers. Blacks are different - the WEAKEST students apply disproportionately (it’s more than simply that black students are pretty weak as a group, but their weakness does not deter them from applying. The reason for this is twofold: 1)Are black students deluded or misled and think they are stronger than they are? Do they have no real understanding of just how weak they are as candidates for a school like Harvard? 2) Harvard cynically markets to them incessantly because they want to increase applications from this group. (I recently walked through Cambridge and saw a staggering number of tours–largely black students–it just struck us). Is Harvard making all these kids waste their time, just so they can say we don’t favor blacks over whites and Asians, see their “admit rates” are basically identical. What they fail to mention of course is that 50% of their black applicants (at least the bottom two deciles) are INSTANT rejects. This is of course another subject --strategy to engineer outcomes.

Take a look at the discussion of AI here:https://www.soccernation.com/the-ivy-leagues-secret-academic-index-for-recruiting/

2049

@collegemomjam

UChicago surely wants to increase the number of applications from the most desirable URMs, who currently apply to HYPS and this will help do the trick. However, UChicago is NOT going to accept more applicants with lower SES, despite what it proclaims. It simply doesn’t have the resources to compete with HYPS in that regard. It wants these high need students to apply but unlikely to admit them unless they have other hooks that make them more desirable from UChicago’s perspective.

2050

To summarize -or rather distill–as much for myself as anyone else:

Just to provide a little context to the z-score (standard deviation) data that has been released in the litigation, for people who are not used to thinking about standard deviations -and how they are applied.

Importantly, no one is saying that ALL black kids at Harvard are admitted because of their race. Harvard’s OWN estimates show that IF Harvard admission was all about academics, then only about 0.7% of the admits would be black. However, those estimate also show that under the holistic process – if RACE did not receive preference but all the OTHER attributes like ECs, essays, personal qualities, etc. did – then about 2.5% of the admits would be black, a big jump. With race preferences, black representation jumps almost fivefold, which means about 20% of the actual black admits are FULLY QUALIFIED under holistic criteria. So, that is the amount we would expect to see with comparable stats to whites and Asians.

I’m told that it is easy to show from Tables B.3.1 and B.3.2 the following selected points:

When legacies and development admits are excluded, 87% of white admits are above the AVERAGE black admit on the academic index, which is basically a composite of SAT (or ACT) + GPA + SAT2, and 97% of Asian admits are above. (This is from the baseline data in B.3.1R.)

Using the SD of the black admits, one can see that, nevertheless, 20% of the black admits are at or above the white admit average, while only 13% are at or above the Asian admit average.

In the absence of race preference, but including all other holistic criteria, we would expect to see all these numbers at or around 50% - no racial group would be significantly above any other racial or ethnic group.

Adding legacies and development (the extended data set in Table B.3.2R), only 80% of white admits are above the AVERAGE black admit on the academic index, while 94% of the Asian admits are above. (To me, this proves definitively that the legacy and development groups are lower in academic stats, as these preferences overwhelmingly favor white students and not blacks, but I would welcome alternate explanations.) Again, it is important to note is that only athletes are held to a minimum AI. No other group has to meet that minimum.

Again, as above, using the SD of the black admits for this expanded dataset, 21% of the black admits were above the white average, while only 10% were above the Asian average.

It’s easy to go through the same exercise for all the components of the Academic Index like SAT or GPA, but I will leave that to someone who is better with Excel than me! I would welcome anyone pointing out any mistakes in this analysis, but these figures would suggest that under the current holistic system – which include large bumps for race – the black admits are quite a bit less accomplished on academic measures as the other groups. Again, only about 20% of total applicants would have been admitted without the race preference according to Harvard’s OWN ESTIMATES, and the analysis above I’d suggest is consistent with that.

2051

The 3 students were not selected based on SAT score. They were selected based on a comment in their summary sheet. One can estimate the rate of that comment appearing on the summary sheet with good accuracy because a large number of samples were checked to see whether that comment existed.

However, the summary sheets were not selected based on SAT score. This results in little precision and meaning when taking the average score of those 3 students. We don’t know whether a different set of 3 students would have completely different SAT scores or not. What we do know, is the average scores of those 3 applicants were similar to scores of typical rejected Black applicants and no where near the scores of typical accepted Black applicants.

The documents don’t say anything about them being wait listed. instead they say 15% of Asian applicants were in this group, and 1% of Black applicants were in this group. This doesn’t sound like wait list rares to me.

Across the full expanded sample , 92% of Black applicants were rejected and 8% were accepted. Figure 1.2 indicates that the average score of all Black applicants was ~1240, and the average score of the few Black applicants who were accepted was in the low 1400s. Among the group with average scores of ~1240, 92% were rejected. Hardly any Black students get accepted from the group with average scores this low, the few that are accepted average tremendously higher scores.

You asked about the SD rates compared to the admitted students in your earlier post for all races. Numbers are below for an average of all sampled years. The documents suggest a mean SAT section score of ~740 and SD of ~70 for the sample period and mean of ~750 and SD of ~60 in 2017 full class.

Unhooked (besides URM) Sample
Asian Admits: +0.20 SD on SATV, +0.46 SD on SATM, +0.09 SD on GPA
White Admits: +0.16 SD on SATV, +0.09 SD on SATM, +0.05 SD on GPA
Hispanic Admits: -0.40 SD on SATV, -0.40 SD on SATM, -0.04 SD on GPA
Black Admits: -0.46 SD on SATV, -0.65 SD on SATM, -0.28 SD on GPA

Full Sample
Asian Admits: ~770 SAT section score in 2017, +0.20 SD on SATV, +0.44 SD on SATM, +0.15 SD on GPA
White Admits: ~745 SAT section score in 2017, +0.12 SD on SATV, +0.04 SD on SATM, -0.05 SD on GPA
Hispanic Admits: ~730 SAT section score in 2017, -0.37 SD on SATV, -0.39 SD on SATM, +0.05 SD on GPA
Black Admits: ~720 SAT section score in 2017, -0.45 SD on SATV, -0.65 SD on SATM, -0.18 SD on GPA

2052

A Washington Post article from a Harvard Alum interviewer, whose personal experience supports the SFFA findings.

https://www.washingtonpost.com/local/education/harvard-interviewer-confesses-i-think-my-school-had-anti-asian-bias/2018/06/24/56dd0af2-74e8-11e8-805c-4b67019fcfe4_story.html

2053

I usually like to look at old posts after new revelations come out and see if my critical thinking is robust enough to anticipate the outcome with some degree of accuracy. I am pleased that the results are not inconsistent with my thinking all along. Why then some posters who are as intelligent or more intelligent than I am be so wrong? It can not be critical thinking, can it?

I think it is the fear of stereotyping. Here is an excellent piece on this important topic by someone who really knows what he is talking about, and shows great critical thinking to boot:

https://aeon.co/essays/truth-lies-and-stereotypes-when-scientists-ignore-evidence

2054

@hebegebe thanks for that link to that interesting article.

My main take away from that article isn’t so much that the author supports the SFFA’s findings (which he obviously does based on his personal and real experiences), but is a point that I have tried to make on this thread a few times. So while I am totally empathetic to the fact that so many Asians (and some others too) with great stats are turned down while others with lower stats are given spots, these “rejects” are still making their way. They are landing on their feet. There is no way of ever knowing or proving it, but are they truly worse off because they didn’t get into Harvard, or in his wife’s example, Cal Berkeley?

I get both sides of this argument. And we can never make everyone happy. There just aren’t enough spots. Harvard may lose and we may see more Asians at the most elite colleges as a result of this. It really doesn’t impact me or my family because I don’t see my son (my girls are already in college, and one got into an Ivy and chose not to attend) seriously thinking he has a shot at an Ivy (despite having great stats, because as the article says, it’s a crap shoot). But some people better be careful what they wish for because if Harvard loses, the results might not be what you wanted.

I’m not sure what the real gripe is either…is it that colleges employ “holistic” admissions criteria at all (something I fully support for a bunch of reasons) or is the gripe that some think they should be more honest and just admit they are trying to control the racial make up of their classes? (Which I feel in some ways they have…yet they have this weird point system that seems like it’s trying to justify their decisions?)

2055

https://www.nytimes.com/2018/06/25/opinion/harvard-asian-american-racism.html?action=click&pgtype=Homepage&clickSource=story-heading&module=opinion-c-col-left-region&region=opinion-c-col-left-region&WT.nav=opinion-c-col-left-region

2056

@collegemomjam

For some people, I think the gripe is indeed against holistic admissions generally.

Others support holistic admissions, but within limits that they believe Harvard is exceeding. Whatever else it may be, Harvard is still primarily an academic institution. Presumably, even most of the most fervent supporters of holistic admissions would concede that there are any number of fantastic and accomplished teens who would and should not be good candidates for admission to an elite university - i.e, an award-winning student musician with tons of leadership in a volunteer activity with a combined 900 SAT and 2.8 GPA. Obviously, the academic disparities under discussion here are far less extreme, but the principle is the same: one may “gripe” if they feel that numbers reveal an over-reliance on the “holistic.”

This becomes particularly problematic for those who believe that this over-reliance is either deliberately discriminatory or has a disparate impact on one particular demographic group, especially in the historical context of universities adopting holistic admissions precisely as a means of discriminating against a different demographic group.

But beyond these more philosophical problems with holistic admissions, there’s also a potential problem with the specifics of Harvard’s policy - namely, the existence of a highly arbitrary “personal” score that seems heavily and inexplicably tilted against Asians and in favor of URMs. This practice would seem to go beyond holistic review; one can believe that Harvard can and should take into account extracurricular activities and institutional needs (including diversity) while viewing an amorphous personality assessment that has the effect, if not the intention, of disadvantaging Asian students relative to their academic AND extracurricular profiles as at least highly suspicious.

Maybe some of this could be solved, as you suggest, by Harvard simply dispensing with the personal score and admitting that they are making adjustments for demographic balance. The problem for Harvard is that at a certain point, whether you have the fig-leaf of the “personal score” or not, it becomes very hard to argue that they aren’t effectively imposing a quota system in which the demographic percentages are designed to remain stable year by year regardless of changes to the applicant pool. And that’s illegal.

2057

There are two interrelated gripes.

  1. Quotas are illegal, and the argument is Harvard has created a quota system in all but name.
  2. In order to create an illegal quota system, but make it appear not to be a quota system, adcoms have discriminated against Asians in how they are rated in personal qualities.
2058

Interesting assertion that socioeconomic admission could produce the same results.

https://slate.com/news-and-politics/2018/06/harvard-should-use-socio-economic-not-racial-preferences-in-admissions.html

2059

Thanks for the replies.

I think overall we have determined that generally speaking even the URM’s that are admitted to Harvard have strong stats. So do we need a floor score or GPA that makes it not discriminatory to select a black applicant over an Asian? If students with 1400 SAT’s have a track record for getting through Harvard just fine, why is it discriminatory to chose them over someone with a higher score?

If you are hiring a waitress and someone has 10 years experience and someone has 5 years experience but you feel that the applicant with less experience is a better fit or your restaurant, is that discriminatory?

At the risk of sounding like I really don’t understand the laws here (which I truly don’t), why is it illegal to have quotas or at least demographic targets? I’m not giving my opinion either way, just truly asking why is this illegal?

Is it illegal for a broadway show director to select a black applicant over a white applicant to play a black person in a play?

I just feel like we are trying to put numeric values on things that maybe we just cannot.

Again, I’m really not sure how I feel about the whole process. I’m just not sure I disagree that Harvard should be able to pick who they want.

Why can a bakery decide not to make a cake for a gay couple? Where do we draw the line on what discretion private institutions can use?

Discrimination itself is so subjective…and when you protect the rights of one party, are you discriminating against the other? Because it’s Harvard are we holding them to a higher or different standard??? I guess whatever new laws come out of this will apply to everyone, but I doubt anyone will be under the microscope like Harvard.

2060

If you are interested, the relevant opinions to read are Grutter v. Bollinger, and the two Fisher v. University of Texas Supreme Court decisions. As to your other comments, those are drifting outside the scope of the thread, and bordering on the forum prohibition against discussing politics.