There are many colleges that run a delicate balancing act with their finances and may suffer severe consequences, if their tuition revenue decreases by admitting too few full pay kids. However, I don’t think Williams is one of these colleges. Williams has a huge $1.4 million per student endowment. Tuition only composes a small fraction of their revenue. Some specific numbers are below. Drexel appears highly dependent on tuition, with their relatively small endowment and lack of other income sources, but not so much Williams and the others.
% Core Revenue Sources – Median of 2018, 2019, and 2020
Tuition + Fees: Princeton = 6%, Stanford = 7%, Williams = 18%, Drexel = 64%
Gifts/Grants/Contracts: Princeton = 27%, Stanford = 38%, Williams = 17%, Drexel = 17%
Investment Returns: Princeton = 65%, Stanford = 48%, Williams = 64%, Drexel = 5%
As others noted upthread, the Princeton and Stanford numbers are difficult to translate into undergraduate per-pupil spending because of a hospital system, nine-digits’ worth of annual federal funding, restricted philanthropic funding of graduate research, etc.
Williams is effectively an entirely undergraduate institution. Per their own publications, student charges make up 33% of revenues. Despite the endowment’s providing 48% of the revenue, they still rely on an enormous amount of non-endowment giving to balance the budget each year.
Of course Williams and peer high-per-capita-endowment peers are less undergraduate-tuition-dependent than someplace like Drexel. That doesn’t mean they don’t desperately need tuition money, though, and it goes a long way toward explaining how skewed their student populations are toward upper-class families.
Tuition certainly still is an important source of revenue for these well-endowed colleges. However, having their student populations continue to skew toward upper-middle class isn’t necessarily an indication that these colleges “desperately need tuition money”. MIT/Calech, for example, have plenty of other sources of revenues, are need-blind and meeting full need, give no preferences to donors/legacies, etc., are arguably more meritocratic than other private US colleges, and can probably survive without charging any tuitions. Yet, their student bodies still skew toward upper-middle class.
This is why the top 30 or 40 schools in the country are so coveted. Even for rich people the sticker price is starting to be unpalatable for middle and lower ranked schools. So the application race to get Johnny into Duke or Amherst (he may not have a chance at Ivies anymore - people are becoming more aware) has increased. Rich people are still willing to pay $320k for some of these lower top schools, but the over 40 rank schools have to play alot of games with merit and athletics in order to discount enough full pays to meet budgets, while still accepting full full pays and giving financial need based scholarships.
I’m guessing we are approaching the point where rich people won’t want to pay sticker for top 25-40 anymore (sticker is just too high) and those names are just “not worth it”… so some schools that did not do merit before are now quietly adding merit scholarships to their toolbox (aka discounts to sticker).
Back to Amherst, by forgoing legacy preferences, they can give a boost to (perhaps better qualified) non legacy full pay and non full pay kids
I doubt non-legacy admits will be better qualified as I don’t think Amherst is letting in unqualified legacies. They have far more qualified applicants than they can take. They’ll simply find another wholistic factor to use to break ties between their many unhooked applicants instead of legacy.
But Amherst also did something far more important that will actually result in more lower SES families being able to send their kid—they increased grant money:
Amherst will be expanding its no-loan financial aid program, providing full-tuition grants to those whose families earn in the lower 80 percent of U.S. household incomes, increasing aid for families earning up to $200,000 annually, and capping expected work-study contributions at four hours per week., and capping expected work-study contributions at four hours per week.
“ncreasing aid for families earning up to $200,000 annually”
Discounting for the non poor middle class here. They can’t afford 75k a year and many won’t qualify for financial aid at most colleges -looks like a combined family income of 150k gets a discount. 150k in most parts of the country is upper middle class and Amherst is probably trying to find a way to let in those qualified kids in the financial “donut hole” at most colleges.
There is a reason why Drexel is need aware, and Williams + the listed other high endowment per student colleges are need blind. One factor is their high endowment per student makes them far less dependent on tuition income. Another factor is as 1NJParent mentioned, their student population naturally skews towards wealthy kids. Wealthy kids are far more likely to apply to selective private LACs, far more likely to have the scores and other factors that are associated with admission, far more likely to be athletes (more than 1/3 of students at Williams/Amherst are athletes), etc. Legacy preferences also help boost portion of full pay kids, although Williams/Amherst is not as dependent on these legacy preferences as various other schools, for the reasons listed above.
Interesting about Olin, I didn’t realize it was quite so new. In the case of HMC, not sure sibling preferences would explain it since that would still have to reported under California law, and HMC didn’t so report.
It’s possible that HMC hasn’t taken legacy status into consideration for a while but doesn’t want to make an annoucement yet so it can determine the impact without committing itself. JHU supposedly stopped preferential legacy admissions since 2014 but didn’t make the annoucement until a couple of years ago.
Having just finished helping my second son navigate the selective school application process, it seems a lot depends on winning the parent lottery. Kids’ exposure to high quality K-12 education is unfortunately tied for the most part to where the parent(s) bought their homes. And how much they saved up. And how much they are savvy about college admissions. And having enough time to help with driving to EC’s. If divorced or never married, parents need to be at least civil enough with each other to fill out the fin aid forms. And so on. These universities really try hard to be more economically diverse, but there are just so many barriers to entry.
Yes, I agree that wealth has impact on kids’ college admissions in many ways. Here’s an article from a few years ago:
It lists a number different roles wealth plays in college admissions, including:
Counselors
Essay “help”
Testing
Where colleges recruit
“Demonstrated interest”
Early decision
Legacy preferences
It’s interesting (and relevant to this thread) that the article mentioned some of the quotes in the documents Inside Higher Ed obtained from US Dept of Education when it investigated Harvard’s practice of legacy preferences:
“Well not much to say here. (Applicant) is a good student, w/average ECs (extracurriculars), standard athletics, middle-of-the-road scores, good support and 2 legacy legs to stand on … Let’s see what alum thinks and how far the H/R (Harvard/Radcliffe) tip will go.”
“Not a great profile but just strong enough #s and grades to get the tip from lineage.”
“Without lineage, there would be little case. With it, we’ll keep looking.”
I guess the last quote is what they meant by “giving an applicant a second look”.
I know that when my kid was approaching college age, Stanford 's policy for non-major donor legacies was “a second look”. They guaranteed that legacy apps would have 2 readers - a bit of protection against a reader in a bad mood, one who had a less favorable reaction to something in your app (style, activities, etc).
@Data10 might know if this is still the policy. It always felt like additional "insurance " for the school too when a parent came back unhappy about junior’s decision. Much easier to say that 2 people had turned down a kid rather than having it all hinge on one reader.
Yes, Stanford is known to give a legacy applicant “a second look” (it’s also in the Stanford Magazine article I linked to earlier in this thread), but my last comment wasn’t really only directed at Stanford. I believe the practice is much more widespread among schools that offer legacy preferences (Harvard is certainly one of them, based on the quotes from Inside Higher Ed above).
However, Stanford’s “second look” may be significant in the following sense: Stanford seems to do everything possible not to give a typical applicant “a second look”. Stanford’s standard practice is to reject most of its REA applicants, rather than to defer them, and to give them “a second look”, in the RD round, unlike some of its peers. MIT similarly gives its EA applicants little extra boost, but accepts a relatively significant portion of them in the RD round.
I wonder about the mechanics of the legacy “boost”. I’m very skeptical of the “second look” theory.
My guess is that the boost is not some second look and/or thumb on the scale (or whole arm as the case may be at certain schools). I’m guessing each school first targets a rough percentage of the class for legacy admits, and then legacy applicants largely compete only against other legacy applicants.
The Harvard litigation data and analyses are compelling as regards legacy preference. There is no question in my opinion that legacy admits at Harvard are weaker on average than unhooked admits, at least on the metrics Harvard cares about. Why wouldn’t it be the same at Amherst and all the other super selectives that use the hook?
Objectively, that relative weakness is probably pretty small, reflecting the reality that there are many very similarly qualified applicants in the applicant pool as a whole. But the admissions advantage for legacy, conversely, is huge. Legacy applicants are not running in the same race.
I would agree that legacy admits as a whole, 20+ years ago were probably weaker than the other unhooked admits when a larger portion of each entering class were legacies. I would also agree that the “donor (min 7 figures)” kids or kids of someone “famous/powerful” are also likely to be academically weaker. More recently though, I think the advantage is more in the nature of a potentially “soft quota” of legacy admits that you mention. If we look at @Data10 post in 162 above, we can see that White ORM LDC applicants have generally higher Academic ratings (54% have a 1-2 rating while only 45% of White non-ALDC are 1-2). This is consistent with what the Yale AO has said on various occasions to interviewers and alums whose kids are applying. The advantage is if out of 100 admits of 2,000 applicants, 12-15 spots are set aside for legacies, and if the number of legacy applicants is say 50, the admit rate for legacies would be 24-30% vs the overall admit rate of 5%. It could well mean in fact that the average accepted legacy has higher academic qualifications than the average admit because the competitive pool is much stronger.
The average legacy admit does have higher qualifications than the average admit, at least in the Harvard data. But the average AA admit does also, as well as the average Hispanic admit and Child of Staff admit. In fact, just about every category has higher than average stats. This is because recruited athletes bring down the average so much that everyone else looks good in comparison. Not picking on recruited athletes here, just pointing out what the data says.
The average legacy is without a doubt less qualified than the average non-legacy (absent other hooks). Otherwise there would be no need for Legacy preference - they would get in on their own without it.
I tend to look at academic qualifications of admitted students in terms of their distributions, rather than averages. If unhooked students form a typical bell-shaped distribution, the distribution of legacy students is likely to skew to the right (i.e. more qualified on average) but with a fatter tail on the left (meaning that there’re greater percentage of relatively less qualified students). The distribution of athletic recruits skews to the left (i.e. less qualified on average) but still has a tail on the right side (meaning that there’re some percentage of these recruits who are highly qualified). Similarly for all other groups. Their distributions are all different but they all overlap to some degree.
Rather than looking at portion of applicants with different academic ratings, it may be more meaningful to compare admit rate for those legacy and non-legacy applicants with a constant academic metric. For example, the admit rate by applicant Academic Index decile and Academic Rating is below for LDC vs non LDC. The admit rate for non-ALDC kids in top 3 Academic Index deciles was similar to the admit rate for LDC kids in the bottom 3 Academic Index deciles. The analysis found that the average boost was ~10x for legacy and 22x for double legacy, after controls for academic qualifications of applicants, hooks, REA vs RD, etc.
Harvard Admit Rate for White ORM Applicants by AI (decile is reference to full applicant pool)
Top AI Decile: LDC Admit Rate = 57%, Non-ALDC Admit Rate = 15%
2nd AI Decile: LDC Admit Rate = 57%, Non-ALDC Admit Rate = 11%
3rd AI Decile: LDC Admit Rate = 47.5%, Non-ALDC Admit Rate = 7.5%
4th AI Decile: LDC Admit Rate = 36%, Non-ALDC Admit Rate = 5%
5th AI Decile: LDC Admit Rate = 32%, Non-ALDC Admit Rate = 4%
6th AI Decile: LDC Admit Rate = 26%, Non-ALDC Admit Rate = 2.5%
7th AI Decile: LDC Admit Rate = 23%, Non-ALDC Admit Rate = 2%
8th AI Decile: LDC Admit Rate = 17%, Non-ALDC Admit Rate = 0.5%
9th AI Decile: LDC Admit Rate = 12%, Non-ALDC Admit Rate = ~0%
Bottom AI Decile: LDC Admit Rate = 6%, Non-ALDC Admit Rate = 0%
Harvard Admit Rate by Academic Rating
Academic Rating = 1: 97% Admit Rate for LDC, 68% for non-ALDC
Academic Rating = 2: 49% Admit Rate for LDC, 10% for non-ALDC
Academic Rating = 3: 18% Admit Rate for LDC, 2% for non-ALDC
Academic Rating = 4: 3.5% Admit Rate for LDC, 0% for non-ALDC
Academic Rating = 5: 0% Admit Rate for LDC, 0% for non-ALDC
I was not addressing whether or not legacy gives a boost at Harvard. It certainly does and a strong one. The boost may come from a soft quota where legacies are competing for a proportionately larger number of spots relative to the number of legacy applicants vs the general pool, using my example. Or the boost could come from lower standards for legacies. Or, more likely a combination of both. The question I was addressing was whether the avg legacy admit had a lower academic qualification/rating than the general pool. Looking at the Arcidiacono ALDC report Table D2, it appears within the 4 race groups, there are a smaller percentage of LDC admits in the Academic 1-2 group vs non ALDC, with the exception of Hispanics where the LDC percentage is in fact higher:
White
African American
Hispanic
Asian American
Typical
LDC
Typical
LDC
Typical
LDC
Typical
LDC
2 or better
88.77
78.34
59.39
43.21
65.4
70.49
94.4
85.56
3
11.19
20.93
40.52
56.79
34.6
29.51
5.6
14.44
4 or worse
0.04
0.73
0.08
0
0
0
0
0
So this does mean within race groups, there are a lower proportion of higher academic rated kids in the LDC group (as a group less qualified) vs the non-ALDC group (other than Hispanics) for this period at Harvard. How much of this is driven by the DC’s I am not sure, but the L’s do make up most of the LDC’s.