According to Parchment, University of Florida does better against Harvard than either Yale or Princeton.
One shoulds not assume that HYP is the top choice among all cultural/political/religious/economic demographics.
“What does Parchment have to do with anything.”
This entire thread is based on Parchment data. See the original post.
JHS: With regard to the fact that students can apply only to one of Oxford and Cambridge, I don’t think a British student loses much by that, because admissions there are much more predictable than admissions at top US schools, and they are very heavily academically based. Here, a student would be well advised to cast a reasonably broad net, to get into a top US school, especially if the student had “only” top-flight academics going for him/her.
Also, I like Mastadon’s post #40.
If a college wants to “win” the revealed preference pool, the college should not admit a student if the admissions staff thinks the student might go elsewhere, if admitted to both schools. I don’t think Harvard worries much about this. But there used to be a strange pattern at Princeton, where the odds of admission went up with combined SAT scores up to a certain (high) point, and then dropped off.
@northwesty - The focus is on UChicago because, not to put too fine a point on it, they’re playing a different game that makes it hard to compare them with anyone at this point.
First, if UChicago were to publish yield numbers that were higher than anyone else’s (as may well be the case) what conclusion would you draw? That they’re the most desirable college in the country? That view just isn’t credible.
Last year, by all appearances, UChicago cherrypicked the vast majority of its class from kids who had committed to attend if admitted, and (again by all appearances) waitlisted many of the other really desirable candidates so that they could then commit and be offered if they were otherwise left out in the cold and UChicago had spots for them. This was unique among peer schools in its scope, as far as I can tell.
Can I prove this? No, but that’s what the circumstantial evidence (and some anecdotal evidence that I’ve observed directly) would suggest. I hope one day UChicago releases the statistics that can prove or disprove the above. If I’ve fundamentally misread the situation, I’ll eat crow until I turn maroon.
I don’t think they want to publish their overall yield number, because few people believe that UChicago, great though it is, is the most desirable school in the country, and a yield number that indicated such a thing would rightly be seen as the byproduct of an admissions policy based around targeting mostly those people who were virtually certain to enroll. It would make it clear that an overwhelming percentage of last year’s class was admitted ED.
The other statistics that I think they don’t want to publish are the ones that will, in my estimate, show that (i) UChicago’s ED admit rate last year was much higher than at schools they believe they’re superior to, like Penn and Columbia, and (ii) its RD admit rate was at or below those of Harvard and Stanford, which relatively few believe they’re superior to.
Why wouldn’t they want this information out there? Because RD apps would crater, and all the work done over the past few years to drum up apps, increase the appearance of selectivity and pump up the US News ranking, all the printed materials that UChicago proverbially sends out to vast numbers of candidates every year, would be for nought.
Regarding the Parchment info, if the majority, or a substantial plurality, of responses come from California and Michigan, as seems to be the case, I’m going to suggest that UChicago’s cross-admit data might be benefiting from some in-region bias from Michigan kids who want to stay in the Midwest. How many people turn down HYPS for UChicago each year? I don’t think nearly as many as Parchment suggests.
I’m also not sure how to interpret some of the cross-admit data. If you believe Parchment, UChicago cross-admits choose Yale by a higher margin than, in turn, Harvard, Stanford and Princeton. That may well be true, but I don’t see an obvious reason for it.
This thread has veered off on a yield/YTAR tangent. Getting back to Parchment…
I really don’t understand why self-reported data is summarily disregarded by so many folks on CC. The census data are self-reported too. (Nobody comes into your house to verify the number of people living there or how many rooms that you have.) Should we refuse to accept the reapportionment of members in the House of Representatives because the census data doesn’t pass the purity test?
I’m retired now, but worked for 27 years in a business in which it was necessary to meet hard deadlines for billion dollar programs, even when the only available data for a critical phenomenon were of dubious quality, or a “family” of data with only two or three members could be constructed for the phenomenon.
You use what you have available. Place as many caveats on it as you want. But throwing out the only available data because it is self-reported is nonsense.
I should have written:
But throwing out the only available data because it is self-reported or sparse is nonsense.
Bcos this is college confidential, as any student in AP Stats knows, self-reported anything is NOT data – its just a collection of anecdotes.
btw: the census has a very high return rate (mid-90’s), so its results are not as subject to manipulation by need-do-wells. Not to mention the fact that census forms are mostly completed by adults, not teen age kids who just got jilted by their ‘dream’ school.
What is Parchment’s return rate? (hint: it doesn’t have one.)
You can’t have it both ways. If “self-reported anything is NOT data” then the census results are NOT data. Your argument then apparently relies simply on the return rate, not on the fact that it’s self-reported. If the Parchment return rate were high enough, you’d be forced to acknowledge that these self-reported data are valid too.
Other self-reported college results… Surveys on: sexual assault, drug use, alcohol abuse, etc. I guess that conclusions based on those “collections of anecdotes” are bogus too. Nope, can’t come up with educational policies based on these surveys. Out of luck.
I personally think the Parchment data is useful, @whatisyourquest - it’s the only source we have, and far better than nothing. It’s self-reported, sparse, and apparently geographically skewed, though, and those limitations should be recognized.
As an example, here is Parchment’s “Student Choice College Ranking 2018”, which it states is derived from the revealed preferences of 88,000 self-reporting respondents: http://www.parchment.com/blog/press-release/parchments-7th-annual-student-choice-college-rankings-reveals-selections-nearly-88000-students-nationwide/. Harvard comes in at #37, behind, for example, Tufts (whatever happened to “Tufts Syndrome”?), Evangel University in Springfield, MO, and the Art Center College of Design in Pasadena, CA. Really?
@DeepBlue86 I agree. All of those limitations should be acknowledged. But what Parchment has is, as you said, “far better than nothing.”
I should have included ethnicity and sexual orientation in my list of college self-reported survey results. You check the box that you think fits best. It’s ludicrous to expect validation through DNA tests. These self-reported data (I’m calling it data) are routinely used by colleges, as well as by the federal and state governments, to set policies and goals. If you are a purist about self-reporting and statistics, you must reject this practice.
Blue – I agree that the ultimate data on this would be cross-admits – if you could actually get all that data. That’s the problem. And the other problem is that the size of the cross-admit pool keeps shrinking. The number of seats getting allocated via SCEA and ED is becoming so large that today’s RD pool is not your father’s RD pool.
It will be interested to see what yield data UChi puts out. Up until now, the distortive effects of ED and SCEA have not messed up the utility of yield/YTAR as a ranking tool. Maybe now you’d have to asterisk UChi.
Although UChi did have near Princeton yield last year, which was without ED and (I believe) without SCEA either. So last year UChi could have argued that they and MIT were the only ones who were not gaming their results.
When you add in the Parchment data (much better than I would have expected) I think you can argue that UChi has made it into the HYPSM club even accounting for their recent ED gambit.
It seems to me that the vast majority of people using SCEA for HYPS will be using it for the first choice school (among that group, at least). Other than legacy, maybe, there’s not much reason to use it for a school that’s not the first choice. I think that probably results in a bunch of kids who would have been cross-admits never finding out if they would have gotten into those other schools.
You could have left your point at the first word, “If”. Since it can and will never be, the rest is speculation.
That’s is my point. Bona fide surveys conducted by health professionals, social scientists, etc., all have limitations, but they seek to minimize them by all kinds of fancy statistical tests.
Parchment does no such thing. It is NOT a survey. Its a collection of anecdotes.
What federal and state policies/goals are based on Parchment “data”? (btw: colleges conduct their own surveys of acceptees that go elsewhere. It is just rare that they publish the results.)
And yes, race/ethnic makeup of the student body is becoming less well known as the number of Decline-to-State increases.
As Jerry Maquire said “It was just a mission statement.” :))
@prof2dad was just having some fun with numbers.
I was trying to show that your statement that “self-reported anything is NOT data” and can’t be used statistically (any AP stats student knows that, you wrote) is entirely inconsistent with the evaluation of social phenomenon. I cited the census, surveys of individuals about their behavior, responses about ethnicity and sexual orientation. All of this self-reported (and unvalidated) information has been used statistically to extract trends. They are treated as useful data, not a collection of anecdotes. This is not debatable. However, I never equated any of these with Parchment. You are. You could have simply agreed that I have a point that the statistical use of self-reported data is common, and still have insisted that Parchment is not useful, for a variety of other reasons.
I’m not convinced of that, @northwesty - UChicago’s yield from the year before last was without official ED, but (by many accounts) with a good number of offers to applicants on the wait list given on the understanding that those offers would be accepted. Meanwhile, the proverbial high volume of mailings from UChicago to potential applicants increased applications and the market’s perception of the school’s selectivity. Through a virtuous circle, that probably increased yield as the brand improved.
Being the highest-ranked team that doesn’t make the playoffs doesn’t mean you’re in the playoffs “club”, and UChicago approaching Princeton’s yield, particularly given the different way UChicago was already playing the game well before last year, doesn’t put them in Princeton’s “club” with HYS either, as I see it.
The Parchment source data for UChicago is summarized here: https://www.parchment.com/u/college/1416-University-of-Chicago/profile. From what I can tell, four of the top five states for applicants admitted to UChicago and submitting data to Parchment, representing some 35% of that pool, were Illinois, Michigan, Wisconsin and Indiana (the top state was California, at 15%). I have to think that, based on in-region bias, this skews UChicago’s cross-admit performance favorably against the Ivies, at least as reflected by Parchment. In my personal bubble (not in a state bordering one of the Great Lakes), I’ve never met a kid who turned down one of HYPS for UChicago.
Fascinating discussion re Chicago here and on other threads. Thanks to the major contributors. I’ve learned a great deal reading your debates.
But I must pose a question about the reliability of Parchment. The site says UChicago admits have an average ACT of 29. Similarly, Parchment reports admits to HYPS have average ACT scores of 25, 28, 27 and 26—all obviously absurdly inaccurate. Do you trust ANYTHING on Parchment?
Blue – That’s the problem with Parchment – it is very partial data and non-representative.
UChi was within 3 points of Princeton’s yield with unrestricted EA and without ED. What do you think their yield would be if they had chosen to go SCEA instead of ED? It would go up right?
What do you think would happen to Princeton’s yield if they got rid of SCEA (aka ED-lite) and went unrestricted EA like UChi? It would go down, right?
Blue - Just asked my son who is on summer break from UChicago. He knows five students who selected UChicago instead of HYPS (three Yale, one Harvard and one Princeton). Most from the West Coast and each had a different reason for the college choice.