It is possible that the percentages are larger than typical enrolled percentages because:
A. These are admit percentages, not enrolled percentages.
B. The percentages may be inclusive (rather than exclusive) of multiracial students. I.e. someone who is black and NH may be counted in both percentages if they are inclusive of multiracial students, even though typical stats (e.g. from College Navigator) would include that student in the multiracial category, but not the black or NH categories. (Inclusive percentages would add up to more than 100% if there are any multiracial students.)
ucbalumnus summarized it well. Harvard’s website intentionally lists the racial percentages in a way to make it appear that URM enrollment is larger than actually occurs by listing admit instead of enrolled, double counting multi-racial students to make totals sum to >100%, and likely doing something unexpected with international students. I expect they choose to list a higher URM enrollment like this for marketing reasons.
If you look at any other reliable source, the URM enrollment will be much lower. For example, the federal database at https://nces.ed.gov/collegenavigator/?q=harvard&s=all&id=166027#enrolmt says Harvard’s URM percentages are 7% Black and 11% Hispanic, and 0% Native American. IPEDS URM percentages match above. A more detailed sumamry by year is below:
Harvard Single Race URM Enrollment
2017 – 7% Black, 11% Hispanic, 0% Native American
2016 – 7% Black, 11% Hispanic, 0% Native American
2015 – 6% Black, 10% Hispanic, 0% Native American
2014 – 6% Black, 10% Hispanic, 0% Native American
2013 – 6% Black, 9% Hispanic, 0% Native American
2012 – 6% Black, 9% Hispanic, 0% Native American
2011 – 6% Black, 8% Hispanic, 0% Native American
2010 – 6% Black, 8% Hispanic, 0% Native American
2009 – 7% Black, 7% Hispanic, 1% Native American
Some selective colleges choose to enhance URM enrollment like this on their websites, while others do not. For example, Princeton lists the same URM enrollment on their website as they do in the federal databases above. Specific numbers are below, which seem very similar to Harvard.
When you say “single race” does that mean that the student chooses one race to identify with? So, someone who is Black and Hispanic would identify or be identified as either Black or Hispanic but not mixed?
It means that a category like “black” includes only those who checked only “black”, but not those who checked “black” and some other category like “white” or “Asian” (those who checked more than one category are listed as “multiracial”).
Note that “Latino” / “Hispanic” is commonly a separate question, so that (in exclusive categorization) anyone who says “yes” to that question is listed as “Latino” / “Hispanic” regardless of what race categories s/he indicated.
It would give a better picture if both inclusive and exclusive percentages for each group were listed.
I think that nearly everyone would agree with race potentially justifying a slightly lower bar. The issue is whether that bar is in fact SLIGHTLY lower.
From the Harvard lawsuit, Arcidiacono rebuttal report, Table 5.2R:
Academic index decile 5: African A acceptance rate 22.41% vs Asian A acceptance rate 1.86%
Academic index decile 6: African A acceptance rate 29.72% vs Asian A acceptance rate 2.49%
Academic index decile 7: African A acceptance rate 41.12% vs Asian A acceptance rate 3.98%
The academic bar is not very high with an academic index of decile 7 or below. It is questionable that this level would represent decent stats.
Like @hebegebe, I’m all in for outreach programs to find talented URM that might not have considered an elite college, or any college, and I favor a low-SES bump as well. What’s telling about the current status of AA is the fact that there is no bump for low-SES compared with high-SES african americans. Given this type of data, the optics doesn’t look good for holistic admissions.
My post above was misleading. The common app has the following 3 relevant questions for federal race methodology. The single race part only applies to the 3rd question. IPEDS classifies any domestic student who says yes to question #2 as Hispanic, regardless of their answer to question #3. So a student who is Hispanic + Black is classified as Hispanic and not Black, while a student who is White + Black is classified as “Two or More Races”.
Select your citizenship status.
Are you Hispanic or Latino?
Regardless of your answer to the previous question, please indicate how you identify yourself (Select one or more): American Indian, Asian, Black, Pacific Islander, White.
IPEDS New Methodology for Mixed Race
IF (Non-citizen is selected), race = Non-resident Alien
ELSE IF (Hispanic is selected), race = Hispanic
ELSE race = Two or More Races
The Arcidiacono document has a graph showing the specific scores of admitted students. In the most recent available year, the graph suggests the overall admitted class had a mean SAT subscore of slightly over 740 (start of ~99th percentile), while African American admits had a mean of slightly under 720 (start of ~98th percentile)… This seems reasonably consistent with the CDS for that year, which lists a 25th percentile score of 700/710 or 32 ACT. I’d consider a 98th percentile SAT score in the low 700s to be “represent decent stats,” but I realize some others would not. I do agree that African American applicants get a strong boost in admission and are far more likely to be admitted with stats in the range you listed than typical unhooked ORMs.
The issue is not unprepared, unqualified admits but the systematic discrimination against Asians where they are limited because of their race, compared to each other, and need higher scores/gpa, stronger interviews, ECs to get in. Harvard has been using soft quotas for a while now, it will be up to the judge or supreme court to figure out if it has used hard quotas.
It gives them plausible deniability for the discrimination.
Under AA it is vaguely legal to do some kind of racial balancing. You could imagine that if there were a federal law against AA like California Prop 209, the data in admissions disparity that SFFA found would surely be a smoking gun in violation of anti discrimination laws. Most schools which proudly practice AA have much lower Asian percentages than comparable caliber schools which don’t, so that points to the conclusion not just a Harvard specific problem but rather AA in general.
AA also allows schools to achieve “racial diversity” without getting rid of programs that go a long way towards limiting it, like legacy admissions.
The Arcidiacono report states that the mean SAT score (across all sections) from the year 2000 to 2017 for African-American admits was 704, whereas it was 767 for Asian-Americans (looking at the scores for many consecutive years rather than just one year paints a fuller picture of what actually happened historically). Hence out of 1600, this would be 1408 vs 1534. These are mean scores so this indicates that a large percentile of African-american admits (>25%) had scores in the 1300s. The SAT is not that demanding of a test. I would argue that a cohort with SATs in the 1300s (out of 1600) would indeed be relatively unprepared (and a bit mismatched) especially for Ivy league math, physics, engineering courses. Of course this does not mean that you couldn’t catch up if you were willing to put in the hard work and you had a great support system. But for perhaps up to 25% of african-american students, I think a mismatch might have been palpable. Hence the much better records of HBCs (compared with ivy league schools) for generating african-american scientists and doctors as @changethegame had noted earlier in this thread. Is ~25% of a cohort large enough to be concerned about mismatched admits? Perhaps not, and if so, feel free to eyeroll away.
If Asian Americans are being discriminated against in the way black and Hispanic and female students were at one time, then it’s wrong and it needs to stop.
That does not negate the need for affirmative action for those who were and are discriminated against.
Go ahead and argue that, but IMO it has very little impact.
1408 is still higher than 94th percentile. As the mean, there will be higher and lower scores, both. Not sure why you focus on the low. But even the low is 87th %ile.
I believe Asian Americans have suffered at least as much discrimination in this country as Hispanics, as a group, but no one is rushing to give them any benefits. In fact, well-intentioned people are calling to defend the side that’s discriminating against them in the name of affirmative action.
Even nominally, affirmative action is meant to address diversity and not to correct for discrimination. But, as someone who works in the university, I suspect the true reason behind today’s version of affirmative action is a lot more sinister even than that.
It’s long been an open secret that wealthy families can contribute to a school to curry favor with admissions officials. But major research universities, even public schools, receive a lot of private funding from sources that they don’t have to disclose. For example, MIT recently had a discussion about whether to continue to accept funding from the Kingdom of Saudi Arabia. With all of the moneyed special interest groups lobbying for certain segments of the population, it’s not that big of a stretch to say that those groups could be applying inappropriate pressure on universities to maintain quotas of their constituencies, without any public disclosure.
It makes sense then that the size of the preference for minority groups roughly lines up according to their political power in this country. The exception is whites, who—because they can’t justify increased admissions by affirmative action, and because there generally aren’t organizations representing their special interests—use other tactics instead, including legacy admissions, the Asian discrimination, “Z-list”, athletics, and the central place of the Greek system.
It might sound like a conspiracy, but if you think about how everyone’s motivations line up, it makes more sense than the stories we’ve been fed about social justice and diversity. Hardly any university tracks what happens to the URM students when they arrive on campus. There are far more resources devoted to recruiting them than there are for ensuring their academic success. In my department, affirmative action hiring decisions are checked and enforced by the office of the president, by people without the right scientific expertise to evaluate. If you think about how even public schools are ratcheting up the number of full-pay international students every year, it’s clear that MONEY is the main driver behind higher education (as with all industries) and this has something strong to do with AA.
The SFFA lawsuit is not only about whether Harvard discriminates against Asians relative to non-Hispanic whites, but whether African Americans and Hispanics receive an admissions boost over Asians.
Well, even under current affirmative action guidelines there are strong limits on the use of race. So the lawsuit is suggesting that Harvard’s current considerations fall outside the acceptable limits (although the Asian discrimination is the main line of attack, as these arguments did not convince the judges in the Fisher cases).
The main issue, though, is the amount of cover that affirmative action gives to these discriminatory practices. The current guidelines leave a lot to the discretion of the universities, and the universities in turn reward that trust by denying everything and withholding all information possible.
There are quite a few differences in Harvard admission since the 1990s when the graph starts (class of 2000). For example, the graph suggests something changed near the class of ~2010, in which the number of URM applicants increased rapidly compared to other races, likely due to a combination of increased marketing to URMs and the Harvard FA Initiative. If we want to look at Harvard’s admission practices in the 1990s prior to these changes, then we might also consider the old Harvard lawsuit which includes a sample from the 1980s and 1990s. However, if we want to paint a picture of more current admission practices, I think it’s better to use more recent years. The most recent year in the graph was the class of 2017. This was not an anomalous year. Instead the class of 2017 fits well with the trend line of an increase in scores over time as the college becomes more selective, so I’d expect the gap between AA and the overall admitted class to further narrow in the current class than occurs in the 720- vs 740+ point gap per section that occurs in the class of 2017, perhaps being a gap of ~20 points per section today.
Harvard’ s CDS indicates that under 7% of those submitting ACT had an ACT composite below the equivalent of 1370. These 7% of the class with below 1370 were almost certainly primarily recruited athletes and not primarily non-athlete URMs. Recall from the lawsuit that being a recruited athlete was orders of magnitude a stronger hook than anything else including the URM hooks, with odds ratios near 3000x. Athletes with a 4 academic rating had a 70% admit rate, while non-athletes had a 0.07% admit rate – ~1000x lower.
Ignoring the specific numbers, suppose 25% of URM admits do have scores, grades, HS courses, and rest of application suggest they are likely to do poorly in engineering. How do we know that those 25% are interested in engineering and are not instead planning on a more liberal arts focused field? I’d expect Harvard to consider the scores and rest of the application in the context of the proposed major. For example, a 1350 composed of 800 Math and 550 Verbal probably is more likely to be admitted as an engineering major than one who scored 550 Math and 800 Verbal. However, I doubt that either is likely to be admitted unless a recruited athlete and/or there is something really unique in the rest of the application beyond their scores.