"Race" in College Applications FAQ & Discussion 12

Applying to College
2061

It annoys me when news articles misrepresent information with seemingly shocking headlines and statements like this.

If a “terrible personality” is defined as personal rating in the lower half of the 1-6 scale (1 is best, 6 is worst), then only 0.48% of unhooked Asian applicants received this rating, which is very close to the overall average of 0.46% of applicants in the baseline sample. If you instead define “terrible” as meaning not among the highest scoring applicants who are most likely to get admitted, then you might look at the top 1-2 ratings, like the lawsuit has done. 19.5% of all applicants in the baseline sample received a high score of 1-2 in the personal rating. 17.7% of unhooked Asian applicants received a 1-2 – a difference of 1.8 percentage points. The difference in average personal rating between Asian and White applicants was ~0.13 on the 1-6 scale. Both differences are statistically significant, but it does not indicate Harvard “consistently” gave Asian applicants low ratings. Instead it indicates a small difference in the portion that receive a high personal rating, which will likely influence decisions for a minority of Asian applicants. Harvard’s internal review came to a similar conclusion and found the Asian percentage of class would decrease from 31% to 26% by changing from model 2 (academics + legacy + athlete) to model 3 (model 2 + personal + EC).

While Harvard did not indicate Asians “have terrible personalities” and did not not “consistently” give Asian applicants low personal ratings, the personal ratings may be suspicious. The lawsuit mentions alumni rated Asian and White applicants similarly. with 50.2% and 49.8% of receiving the high 1-2 personal rating respectively, There was no significant difference in personal rating with alumni raters, yet a 3.6 percentage point difference occurred with the Harvard raters.

The lawsuit emphasizes what happens when you compare specific academic deciles, which relate to academic stat ranges. Asian applicants have the highest average stats, yet they don’t have the highest personal ratings among Harvard raters. URMs have the lowest stat ranges, but the they don’t have lowest personal ratings among Harvard raters. If you compare the personal personal rating between the typical decile of Asian applicant to the personal rating of the minority of applicants from other races having stats/deciles that high, both alumni and Harvard are more likely to rate the more rare non-Asian top stat applicant higher in the “personal” category, but to a greater extent by Harvard. The specific differences between the share of White and Asian receiving high 1-2 ratings for different academic deciles is below. A higher number means a higher percentage of White applicants received 1-2 ratings. Note that a larger portion of Alumni gave high 1-2 ratings among all races than Harvard, so if expressed as ratios the relative differences would have looked larger.

Share White 1-2 Personal Rating - Share Asian 1-2 Personal Rating
Top Academic Decile: +1% Alumni, +7% Harvard
2nd Academic Decile: +5% Alumni, +7% Harvard
3rd Academic Decile: +4% Alumni, +8% Harvard
4th Academic Decile: +2% Alumni, +4% Harvard
5th Academic Decile: +5% Alumni, +6% Harvard
6th Academic Decile: +4% Alumni, +5% Harvard
7th Academic Decile: +4% Alumni, +4% Harvard
8th Academic Decile: +3% Alumni, +3% Harvard
9th Academic Decile: +1% Alumni, -0% Harvard
Bottom Academic Decile: -2% Alumni, +0% Harvard

2062

@collegemomjam Your question on why quotas would even be illegal is a moral one for me. If a student was not chosen only because of something they can not control (race or sex being the main culprits), it makes for a tough argument. Holistic practices are acceptable when the criteria is transparent (A school like MIT accepting a much higher percentage of women to try and balance out an over abundance of men applicants). But some of the elite schools are not being truthful nor transparent in the admissions practices (data can be skewed but doesn’t lie). One unintended consequence of controversy over these practices is that elite schools are losing qualified URM candidates like my daughter (rising college freshman at a HBCU) and possibly my son (a rising hs junior who got a 1480 on a recent SAT practice exam) because they don’t want to be looked at as being accepted into a school based on being an African American. There is a stigma that has been voiced to some of my daughter’s friends (African Americans accepted into Harvard and Princeton) that they got in because they are black even though they both scored in the top 1% of all college bound students on SAT and were near the top of the class of a high performing high school. Admissions offices showing the criteria used and being honest about their practices should not be a problem, but private institutions with the endowments of the elite schools can almost do whatever they want.

2063

I think the problem with quotas is that if there’s a set number that has to be reached, you essentially have two separate admissions systems. Morally and legally, looking at race as one among many characteristics that may factor into a decision among students of relatively similar qualifications is very different from saying (explicitly or implicitly) that you’ve decided that you want 11% of your students to be African-American, and will lower the floor for applicants in that demographic group as needed.

As far as Harvard is concerned, I’m actually less concerned with admissions for URM students than the apparent artificial depression of admission rates for Asian students. Controlling for legacy, ECs, region etc seems to explain a decent amount of the reason why the percentage of Asian students has plateaued, but the personal score suggests that even beyond that, Harvard is not just tipping the scales in favor of URMs, but against Asians. I don’t think that necessarily suggests conscious animus, but it does, to me, indicate a discomfort with the proportion of, not minority, but WHITE students dropping below a certain number, even though whites would be in no danger of dropping below a critical mass.

Even when it comes to African-American admissions, however, the data raises some questions. Earlier, someone had claimed that the average SAT score for AA students was a 1400 (compared to 1480 for whites); that’s since been changed to a 1440 (compared to 1490), which does make a meaningful difference, IMO, as it suggests that the great majority of black admits are at least in the 1400s. That’s a good score, and I think one could reasonably say that it is getting into the territory of scores high enough that precisely HOW high the score is shouldn’t be the difference.

However, I have a few reservations. One is that SAT score is only one part of what goes into the “academic 1” designation. If we’re talking about an absolutely stellar students for whom a 1420 is actually the weakest part of a very impressive academic profile, that’s one thing. But if in fact we are talking about a student whose GPA, SAT subject tests, and other academic achievements are commensurate with a 1420 - well, that’s a very good student, but someone who, absent a really strong (not just “standard strong”) EC record, is going to need a very heavy thumb on the scale indeed to warrant admission to Harvard.

Especially striking is the fact that the data does not show that African-American applicants are getting a socioeconomic bump; economically disadvantaged black students do less well than economically more advantaged ones in admissions. This suggests that we’re not talking about students whose high but not stratospheric SAT score may indicate nothing more than lack of opportunity for test prep or a shakier educational background. We’re talking about middle-class students. And again, in my experience, middle-class white and Asian students are not generally getting into schools like Harvard with SAT scores in the low 1400 unless there’s something else really special in their application (or they are development, legacy, etc - and I also don’t support lowering admissions standards for these students). This would support the idea that Harvard isn’t just doing race-conscious admissions within a pool of similarly qualified applicants, they’re lowering the floor for certain groups.

The fact that these students are going to be just fine at Harvard isn’t the point. I teach English literature, and certainly think it is an intellectually rigorous discipline. But even as someone who considers myself a pretty tough grader, the bar for passing - and even getting a pretty high pass – in a literature class just isn’t that hard to clear if you’re someone who would be within spitting distance of an elite school. At least in my classes, it is quite difficult to get an “A” range or even a B+. But someone who is getting even a 600 Verbal on the SAT probably knows in a basic sense how to structure an essay, can write with reasonably complex sentence structure and vocabulary, and will be able to manage a banal but on-target “here’s the main theme of this novel, backed up with relevant quotations” kind of paper. That’s not going to be an “A,” but it isn’t going to be a “D” or “F” either. That doesn’t mean that any student who can clear this bar deserves to be at Harvard.

Another factor to consider is that Affirmative Action at Harvard is likely to be far less extreme than AA at other schools. If Harvard - and Yale, and Princeton, and Stanford – are scooping up not only a really high percentage of URMs in the high 1400s and above, but tons of URMs with scores 100 points below that, who is being accepted at the “lesser” elites? I

2064

@ChangeTheGame thanks.

How is it not discriminatory for MIT to admit more females (something you also cannot control) to balance out the class but it is discriminatory to select an African American over an Asian, even if the African American’s scores are “high enough” because they want a more well-rounded diverse class?

My question was more about the URM’s and Whites that still have good stats/scores that are getting spots over the Asians that have higher stats/scores. Is there a floor that makes this acceptable? Should there be?

So I think my question was answered to some extent…maybe it’s just the transparency issue more than anything else that is the problem (I think it is for me). Because I think it’s been determined that many of the URM/White stats are still quite strong, at least as it relates to this Harvard case.

And that’s what I was saying a few pages back…if Harvard could just admit what they are doing and rather focus on why this should be OK and not considered “discriminatory” (to take some URM’s and whites with lesser stats (but still strong enough) than Asians), maybe there wouldn’t be as much of a problem. Isn’t that what happened with the Fisher case in Texas?? Didn’t they ultimately say Texas had the right to do what they did?

As for my point about the waitress being selected with less experience than the other waitress, and the Broadway show actor and cake merchant, these examples are certainly no more political than the Harvard case which is in the courts now. The topic is still discrimination, not which political group supports which stance. Trying to compare when it is OK to take race/ethnicity into consideration or any other potentially conceived “discriminatory” characteristic (like years experience, sexual preference, whatever!!) and when it is not.

Again, I am torn myself. But I’m just not sure Harvard should lose, even though they might.

2065

@ChangeTheGame this makes me sad:

“One unintended consequence of controversy over these practices is that elite schools are losing qualified URM candidates like my daughter (rising college freshman at a HBCU) and possibly my son (a rising hs junior who got a 1480 on a recent SAT practice exam) because they don’t want to be looked at as being accepted into a school based on being an African American. There is a stigma that has been voiced to some of my daughter’s friends (African Americans accepted into Harvard and Princeton) that they got in because they are black even though they both scored in the top 1% of all college bound students on SAT and were near the top of the class of a high performing high school.”

I hope your son will reconsider. Not everyone feels that way and he can choose his friends accordingly. Good luck to your daughter.

2066

I don’t think MIT is particularly transparent about this, and I know that Vassar or Brown or Davidson aren’t in the opposite direction.

MIT:

https://www.washingtonpost.com/news/grade-point/wp/2016/03/13/want-an-edge-in-college-admissions-see-the-schools-where-women-and-men-have-an-advantage/

https://www.washingtonpost.com/posteverything/wp/2015/07/30/achieving-perfect-gender-balance-on-campus-isnt-that-important-ending-private-colleges-affirmative-action-for-men-is/

2067

https://www.insidehighered.com/admissions/views/2018/06/25/harvard-admissions-data-raise-questions-dont-demonstrate-discrimination

Shocking to me that this gets so little attention.

2068

But increasing SES diversity would have real financial costs (more financial aid money needed), while being a less visible achievement in diversity than the more visible race/ethnicity. So it is not surprising that the most selective private colleges design admission procedures and criteria* that result in around half of the students coming from no-financial-aid familieswith top 3% income/wealth, with only a small percentage of Pell grant recipients from families in the lower half of the income/wealth range in the US.

*Note that such criteria do not mean having to be need-aware for individual applicants. But such things as favoring expensive extracurriculars (e.g. expensive sports), requiring NCP information for financial aid, preferring legacies, requiring recommendations, etc. all tilt the playing field toward applicants with the most advantages, without explicit need-awareness.

2069

@OHMomof2 You are right that they could be more transparent at MIT as well. The big difference is there is no drop-off in female students stats in comparison to male students at MIT while there seems to be a statistical significant one that populates some of the URM data presented. I believe that holistic admissions can work, but the data that disturbs me most as a Black man revolves around the percentage of African Americans who finish at the low end of their classes at elite institutions and the percentage of African American students who switch majors from STEM majors to non-STEM majors at mismatch schools (schools where African American students standardized test scores are well below median scores for the the student body as a whole). There may be a happy medium somewhere, but I don’t know if we will find it. Race is treated/perceived differently to legacy status and non-revenue generating athletes and the removal of all would possibly be the fairest solution, but a meritocracy based off of stats alone discounts for the human factor (overcoming great obstacles and odds to succeed) and I believe that should be a part of any metric. But can that be quantified?

2070

@ChangeTheGame

This troubles me as well. I know he gets a lot of hate on this forum, but I believe that Malcolm Gladwell is right to the extent he argues that the very top schools can crush the dreams of some highly intelligent students. The problem is that it takes a good deal of knowledge and strength to turn down Harvard. How many kids have the self awareness to say, “no thank you Harvard, I’m going to choose unknown school XX because I think it will give me a better shot at actually achieving my dreams?”

2071

Aren’t the stats on black vs white acceptances on the order of a couple of SAT questions? I’m seeing low 700s vs high 700s in the Harvard numbers discussed here.

That doesn’t strike me as much of a mismatch.

2072

We are talking about averages. If I were a kid with a low 1400 SAT and I wanted to go into a STEM field, I would think long and hard before choosing a top ten school. If that were my only choice for an affordable education, I might have to take the chance, but otherwise, I would absolutely be looking for some place where I could be closer to the top of the pack.

2073

@gallentjill that’s one way to do it. My kid had a low 1400s SAT (if we define that as below 1450, which I think is actually what it was but her ACT was better), chose a top ten college and is a math major on track to graduate with honors.

Being at the top of the pack has never been her thing, still isn’t. Surrounded by smart, inquisitive students from all over? She does like that, a lot.

But as always, YMMV.

2074

@OHMomof2 Always good to have a counter example. I think my thinking is shaped by my daughter wanting pre-med where graduating in the middle of the pack absolutely won’t do. I also think it may make a difference that her ACT was higher. It would also make a difference how the scores fell out. A kid could get a 36 in math and a 28 in English. So,the statement I was making was very general in nature.

2075

Measures that are correlated to obstacles encountered would be lower parental educational attainment, lower income/wealth, and parental divorce. Race also has some correlation in terms of obstacles caused by racial discrimination. Gender can also have some correlation in terms of obstacles created by gender stereotyping. Legacy correlates to fewer obstacles, since that implies parents with bachelor’s degree or higher attainment. But obviously, there are exceptions, there is a large amount of individual variation, and even the average level of obstacles or lack thereof associated with a given characteristic can vary in different areas, schools, etc…

Grades, test scores, etc. can show how an individual finishes. But it is hard to measure where an individual started or how many hurdles s/he had to jump along the way, even if demographic characteristics suggest that there are significant differences at the demographic group level.

2076

@OHMomof2 The numbers drop very quickly once you look at the pool of African American high scoring standardized test takers. College Board used to release really specific data on the SAT scores by race (15 years ago) and it was a sobering picture (~200 out of 150,000+ African American kids at or above 1450 SAT). I believe that Khan Academy has evened the playing field some and there are more students taking standardized tests than ever, but there are still not many African Americans scoring at a 1400/31 SAT/ACT equivalent level today. They are not splitting out those numbers by race at most schools, but the gap once you look past the top 3-5 schools would start to spread. I know an African American student who is getting ready to go to a top 100 school with scores 250 points below the average student at the institution. Can this student make it? Yes. But it is going to be challenging and there will be an adjustment. The top 10 schools bottom 10% are full of URMs in a similar range below the average student and the same rules apply.

2077

@ChangeTheGame. I find it very sad that your kids don’t want to go to certain schools because of how they may be perceived. My URM kid went to one of the tippy top schools, and she told me years later that some classmates had thought she got in because of AA, but quickly saw that wasn’t the case. She then went to med school, and heard through third parties the same thing.

She graduated #1 in her med school class.

If your kids let other peoples perspective of them define themself, they are putting themselves behind the 8 ball.

2078

There is a reason Gladwell gets a lot of hate on this forum. The reasoning Gladwell gave in the video lecture I saw. was something along the lines of:

  1. A STEM degree is the most important thing a student can have. All students want to be STEM majors.
  2. At both highly selective and less selective colleges, STEM majors tend to have SAT scores higher than non-STEM majors.
  3. Therefore the students who are on the lower end of the SAT score at highly selective colleges all want to be STEM majors, but aren't likely to become one. If they had attended a less selective college where their SAT score was higher than peers, they would likely have been STEM majors.

While Gladwell’s reasoning is nonsense to increase book sales, there still may be an element of truth in his claims. There have been plenty of studies about what happens when students attend colleges with mismatched stats. They usually find little difference in predicted graduation rate or GPA for a given student who is over/under matched. For example, the study at http://journals.sagepub.com/stoken/rbtfl/iAyUJ9v44wd1I/full found

The study that is usually referenced on this forum as evidence of SAT mismatch causing students to drop out of STEM majors is http://public.econ.duke.edu/~psarcidi/grades_4.0.pdf . It found that URMs were more likely drop out of STEM and the rate of attrition could almost fully be explained by differing admission criteria for URMs. The specific regression coefficients for dropping out of numerical majors (engineering, hard sciences, economics) at Duke with full controls are below:

Being Female: +0.188 (0.048) – 99% significant
Year 1 Effect (Degree of Grade Inflation in Specific Classes): -0.084 (0.025) – 99% significant
Admissions Rating of HS Curriculum: -0.147 (0.058) – 95% significant
Admissions Essay Rating: -0.068 (0.044)
Admissions HS Achievement Rating: -0.046 (0.053)
SAT Score: -0.031 (0.027)
…
Being Black: 0.024 (0.064)

Among those who changed majors because the major was too difficult, the coefficients were:
Year 1 Effect (Degree of Grade Inflation in Specific Classes): -0.187 (0.032) – 99% significant
Initial major not STEM: -0.141 (0.071): – 95% significant
Being Female: +0.071 (0.053)
Admissions Personal Rating: +0.076 (0.061)
Admissions LOR Ratings: +0.020 (0.062)
…
SAT Score: -0.009 (0.033)
Being Black: 0.000 (0.078)

The most influential factor by far in dropping a STEM was being female. With the same stats and admission ratings, women were significantly more likely to drop out of STEM at Duke. Few women seem to be dropping out because the classes were difficult. One take away might be to review stats like attrition rate for women engineering at specific colleges. They likely differ for a variety of reasons including things like male/female ratio within the major, role models, and faculty/TA actions.

Relative grade inflation was another key factor. A student could try to avoid classes/colleges that grade harshly in STEM or try to develop more of a tolerance to less than perfect grades. Among students pursuing engineering work after college, getting a near perfect GPA, like HYPSM… type students usually do in HS, is rarely important. Instead most companies that screen GPA use a 3.0 and tend to weight other criteria as much more important for employment decisions. After the first job, few engineering employers care about college GPA.

Among academic criteria, HS curriculum was particularly important. Duke admissions gave the matriculating students at Duke an average rating of 4.7 in HS curriculum on the max 5 scale, making this an incredibly imprecise measure that essentially just flags the minority of students who had below a perfect 5 rating in HS curriculum. Nevertheless, being among the minority with less than perfect HS curriculum rating was the most influential academic criteria in predicting who would drop out of a STEM major. With full controls, SAT score had relatively little influence, as did being Black.

2079

@Data10 Thank you for that thoughtful analysis of Duke’s data. Those statistics where really bothering me and they still do to some extent, but everything is so much clearer and put into proper context.

@CottonTales I think that the perception is a factor for my kids (especially my daughter) because she knows that we didn’t have a problem with that (My wife and I attended HBCUs and had a fabulous experience). But I have always wanted her to make a decision worked for her and I believe she did (there is always Grad School). My son won’t care as much about the perception, and he will apply to some elite institutions, but he hasn’t really gotten to far into the process at this point and could go either way.

2080

@data10 I am certainly not qualified to dig into your statistics. I do believe my statement was more catagorical then I really intended. It probably comes down to the personality of the individual student. I believe there are kids who really do thrive by being pushed by peers who are as or more intelligent then they. They truly want the most rigorous surroundings they can find and will be bored to tears with anything less. Those kids should certainly go to the most rigorous schools they can get into. But, I think there are many more kids who will do better as big fish in smaller ponds. Whether that is because of the curve, or because they are truly smarter then those around them and so stand out, or simply because it gives a psychological advantage…I don’t know.

I personally know of several kids who started out wanting to be doctors and switched out freshmen year after finding the courses too difficult at some very prestigious colleges. Whether they truly decided they would rather not be in medicine or were weeded out of something they might have really enjoyed is something I can’t know. Anecdotes are not evidence of course, but sometimes they are all you really have to make some important decisions.