Espenshade had positive things to say about Hurwitz. But he also understood how a researcher cannot climb into the role of reviewer, deal with the subtleties. We know holistic is more than stats and demographics.
From the Hurwitz study: Using this approach, I eliminate bias in the estimate of legacy status impact due to applicant characteristics that are invariant across the multiple institutions, including those which are challenging or impossible to control for using basic logistic regression (e.g. strength of teacher recommendations, student disposition, and essay quality).
I haven’t read the whole study, yet, but it sounds like he’s attempting to control for way more than just GPA/SAT stats and demographics.
Control for by eliminating, eg, LoRs and essay? You see the limitation?
That’s not it. Hurwitz’s claim is that he can “control for” essays and letters of recommendation by looking at how different colleges treat the same candidate, with the same recommendations and, presumably, the same or similar-quality essays (as well as, of course, the identical GPA, standardized test scores, and demographic profile). It’s pretty persuasive.
LF – unless you hook the adcoms up to a polygraph, you obviously can’t tell exactly why a particular decision was made. But I think you doth protest too much.
When it comes to essays, you wouldn’t think the Harvard legacy is going to be a meaningfully better writer to H as compared to YP. It isn’t that hard for a smart kid to come up with a good “why Princeton” essay even if the kid’s dad went to Harvard. And in the Common App era, HYP are all going to see the exact same main essay.
For rec letters, most kids are going to get/use the exact same teacher rec letter (Jimmy is the best student I’ve had in 25 years) for each of HY and P. While one rec letter for one kid may resonate with one particular adcom, there’s no reason to expect that the teacher rec letters obtained by H legacies across the whole system sing more at H than they do at YP.
Since he had access to the actual files and outcomes, I have to believe that Hurwitz saw many more examples of the pattern where a H legacy got into H but not YP than he did the reverse. And over the entire aggregate data set, that pattern could not be explained by essays or rec letters. Although essays and rec letters could influence one particular decision.
This is four years old, but it would seem to indicate that (at least back then) the largest donors to the School of Arts & Sciences at Penn were likely to be parents of current or former Penn students: https://www.sas.upenn.edu/supporting-sas/sas-supporters/1m-lifetime-donors . It popped up in a Google search, so I assume it’s public.
JHS, but different colleges can be looking for different spin on attrubutes or to fill different institutional needs. There is no one model. Chances reflect the pool. Unfortunately, down to the area competition, others from that hs or down the street, the major, and other factors. One reason the biggies have multiple reviewers is for multiple “takes” on an individual. And the onus is on the applicant, to present that whole that works for each target.
I understand Hurwitz tried to control for certain constants. That H legacy is still a H legacy when Yale reviews her. The competitiveness of her hs is a constant, as is her academic record. Etc. But the actal chance boils down to whether College X likes the whole and who else is a close contender, for a seat at that college, the mix they want to build, for that class.
Most kids struggle, eg, with a Why Us? (That can be deadly, in a close race.) If I protest, it’s because I’m trying to share my own observations. What makes a kid compelling is not mechanical. I can respect Hurwitz’s attempt for what it is. Yes, legacies can end up with (slightly) better chances. For various reasons. And not always. But remember, he was a grad student researcher, trying to observe patterns.
I don’t believe Harvard had a “Why Us” essay when my kid applied. There was a choice of supplementary essays ones of which was a list of books you’d read in the last year which is what my kid do. His main Common App essay was sort of a Why Us essay and he used the exact same one for every school with very minor adjustments. It basically said, “I’m a computer nerd, but there’s still a lot for me to learn. You can teach it to me.” It had a humorous beginning and end, but it was a good for an engineer essay - not a great essay.
I think some legacies may write from a better understanding of the school, but not all of them. And I don’t think it always matters. My kid also told his interviewer that Harvard wasn’t his first choice.
@lookingforward, obviously every kid is an individual, and maybe that Harvard legacy got into Yale in part because he was a bassoonist and Yale needed one that year, but Hurwitz apparently reviewed more than 290,000 applications from more than 133,000 unique applicants, of which 6% were legacies somewhere. That’s about 8,000 kids that he was able to analyze for legacy advantage - so you’d think that the sample size was large enough for the conclusions to be convincing and for the differences in institutional needs to even out. He found the advantage to be clear and substantial, especially if the kid was a primary legacy at one of the most selective schools. What of course he couldn’t fully know or control for was differences in family wealth and giving history among the legacies.
“Marked differences between results emerge from the application of the different analytic strategies, and these differences reinforce the notion that legacy applicants differ from non-legacy applicants across many dimensions relevant to the college admissions process.”
He’s trying to quantify the impact of legacy. Not examine qualitative differences. Even when he discusses “relatively weak applicants,” that’s still based on numerical. I admire the depth of his considerations- and am not disputing that legacies have some advantage. But this study doesn’t tell “why.” It’s just one piece of the vast parts.
Look at the student legacy posts on CC. Some are clearly a different level of thinkers than others, are more aware of what matters beyond stats and standing, and have challenged themselves beyond the usual hs box. For a few, I’ll be interested in their results. But I think you’ve got to agree, @DeepBlue86, that admissions is more than any slice and dice can show.
Of course admissions is more than any slice and dice can show. There’s no formula. On the other hand, it’s pretty powerful when Hurwitz says primary legacies are three times more likely to be admitted. “Three times” may mean an admission rate of 24% vs. 8%, with “only” half the well-qualified legacies being rejected vs. 80% of them for nonlegacies. That’s a big difference, but it doesn’t mean that legacies are shoo-ins, or that they are less qualified if accepted. (Note that my very high-stat legacy kid was part of this data set, and was rejected or waitlisted at all legacy schools.)
And then, on the other other hand, there’s the apocryphal data set at Harvard suggesting that, at least at the tippy top of the food chain, legacy is not really so powerful in itself. Rather, it’s a marker for a certain type of elite insider intellectual and social status.
Not disagreeing on that.
@JHS and @lookingforward - maybe we can agree on the following:
- There is no formula that will tell you ex ante whether any individual applicant is going to be admitted to any particular university (although I might suggest that there's a "Malia Obama exception", where if the daughter of the President of the United States or someone similar is academically qualified and has no personally disqualifying characteristics, it's a virtual certainty she'll be admitted anywhere, because having her "on the team" is overwhelmingly beneficial to any university).
- Legacies of tippy-top universities are, generally speaking, likely to compare favorably to the overall applicant pool of the legacy university and its peers. This is because they likely come from a family that has a strong respect for education and is relatively high-SES, so has probably provided the child with the kind of high-quality educational and other opportunities that make for a strong application.
- Since legacies are likely to be high-SES, it's reasonable to suppose that a significant number of them come from families that are among the most generous in terms of time and money to their universities, which would count in their favor in the admissions process for that university specifically (I think the Penn webpage linked to above at the least doesn't contradict that view).
- Given that legacies are likely to start with the abovementioned advantages, it's not unlikely that some subset of the tippy-top legacy group is going to be extremely strong and not only a good bet to be admitted to the legacy university but also to some of its peers.
- Accordingly, one might expect to find that legacies as a group have a significant statistical advantage in admissions to their legacy university specifically (in line with the Hurwitz study), and some subset of them would be expected to be admitted to multiple peer universities as well (as the elusive Harvard study would seem to indicate).
But just not limited to high SES. It’s not entirely what your parents’ money affords, living in the best school district. It can be the exposure, itself, to a higher order of (dang, that word again,) thinking. And, action. Imagine an applicant whose H educated parents didn’t pursue big money.
We tend to assume the trappings of wealth lead to the “it” and forget the rest of how educated family influences and guides.
“Look at the student legacy posts on CC. Some are clearly a different level of thinkers than others, are more aware of what matters beyond stats and standing, and have challenged themselves beyond the usual hs box.”
Sure – I see those posts too. But those exact same qualities you see in an H legacy will also make those kids stand out at Y and P too.
But when Hurwitz looked at thousands of kids and thousands of admit decisions, he saw a very clear pattern.
Some H legacies also got admitted to Y and/or P. Those could be explained by being an extraordinarily strong applicant.
And some H legacies got rejected by H but got admitted to Y or P. But not that often. Those could certainly be explained by the luck of the draw/roll of the dice/kismet between a specific application and a particular adcom or a particular institutional need.
But most often was that H legacies (after controlling for things as much as possible) fared better at Harvard than they did at Y or P. Like 3-5X better. Going from 10% to 40% at Harvard still isn’t a guarantee and clearly no dummies get in. But the difference is still yuge.
I don’t think we can wave away the conflict between the Hurwitz study and Harvard study that easily. Having a strong application and even having wealthy parents should be absorbed in the applicant specific intercept when doing a CLR. Some possibilities for the conflict:
- The simplest answer is the Harvard study was only about Harvard and Hurwitz was about 30 schools. Perhaps, Harvard is unusual among elite universities in not giving much of a preference to alumni.
- Harvard gives HYP legacies an edge over everyone else. That doesn't seem likely, but it's at least possible.
- HYP legacies tend to be wealthier, and counter-intuitively a donation at Harvard buys more than a donation at other schools.
- There's some methodological flaw in the Harvard study, but we can't determine that without seeing the study.
The Harvard “study” isn’t an academic study, as far as I know. It is the Dean of Admissions (or someone) telling an assistant, “Go back through all the applications this year, and tell me how many applications we got from Yale legacies and Princeton legacies, and what percentage of those applications we accepted, then put it on a piece of paper with the same figures for Harvard legacies. I want to see how different they are.” And the answer is: Some, but not very. Depending on how their summary sheets are set up, it may not even take more than 10-15 minutes of work to do that. Or maybe it takes some work-study kids a couple of days.
It may well be that Harvard legacy applications to Harvard are weaker on average than Yale or Princeton legacy applications to Harvard, since marginal candidates who are Harvard legacies may think that their legacy status gives them a better chance at acceptance, while Yale or Princeton legacies only apply if they know they are solid candidates. The same or close acceptance percentage for Harvard legacies vs. Yale/Princeton legacies may in fact represent a meaningful boost for the (on average, weaker) Harvard legacies.
But . . . I really doubt that the Harvard legacies applying to Harvard are that much weaker than the Yale or Princeton legacies. In my real-life experience, most sophisticated students who apply to any of those colleges (and almost all HYP legacy candidates are sophisticated) do so only if they are decent candidates.
Again, in my real-life experience, it is not rare at all for legacy applicants accepted at Harvard or Yale to be accepted at other similarly selective colleges where they have no legacy status. Provided they apply . . . it is also pretty common for a legacy applicant to apply only to the legacy school (either early or regular), and not to similarly selective competitors.
I’m not sure I see a conflict, @roethlisburger, but I actually see things a little differently from @JHS, for a change.
My guess is that some number of H legacies apply SCEA, get in, and that’s that - they don’t apply to Y, P or anywhere else. That group probably skews strong in the H legacy pool, in that they know enough to apply early (which means they’re savvy and more likely to be high-SES, studies show) and, because they’re applying early, I infer that their families have closer ties to the school and are likely to be big donors. This doesn’t mean that they’re better than the Y and P legacies, just that they’re more likely to get into H and then not apply to Y or P, so their admit rate won’t be compared to that of the Y and P legacies.
Others will apply to H RD (or be deferred after applying SCEA), and, if I’m right and they’re less likely to get into H than the early applicants, is it really so surprising that the the admit rates would be comparable to those of the Y and P legacies, who may be very strong in their own right but don’t have the H donor edge?
^Some of what I wrote above is irrelevant in light of JHS’s statement the “Harvard study” was never intended to be a controlled study. Hurwitz calculated the odds advantage for EA and RD candidates separately. ED candidates were purposely excluded. The odds advantage for EA was greater than for RD, as you might expect. There was still a substantial advantage of 5.5x, even for RD candidates. Hurwitz found a smaller advantage(2x) for what he defined as secondary legacies, those with grandparents, aunts, uncles, or siblings who attended the institution, which was about the same for EA and RD. It’s possible that YP legacies tend to be Harvard “secondary legacies” at higher rates than the general applicant pool.
Neuro-Linguistic Programming
A method of influencing brain behavior through the use of language and other types of communication to enable a person to “re code” the way the brain responds to stimuli and manifest new and better behaviors…
A few hand shakes here, a few hand shakes there. A little bit of fancy dialog and a little money to grease the wheels. An entire generation of siblings admitted to Harvard.
Also, good for dating. B-)