Harvard Cross-Admit Statistics?

<p>Okay, I'm confused. I see so many posters referring to the cross-admit statistics for Harvard, Princeton, Yale, Stanford et al, but I thought all of that information was 1) confidential information not reported by the colleges (perhaps for obvious reasons) and 2) dependent entirely upon voluntary responses from admitted students including both those who chose to attend and those who chose not to attend. </p>

<p>I've seen the link to the Stanford pie chart (posted by Zephyr?) but even it (of course) isn't really a study of cross-admits. For example the 13% (I think it was) who chose MIT over Stanford simply answers the question "where did the students who turned down Stanford end up matriculating?" The answer is that 13% ended up at MIT. In other words, it says nothing about the universe of students who were accepted to both schools and the decisions they made. It could be that 100% of the students admitted to both Stanford and MIT chose MIT and that this group made up 13% of those who turned down Stanford. (I'm sure, however, that this isn't the case.) To have accurate cross-admit figures, you would have to have a series of head-to-head statistics that included all students who were admitted (rather or not they matriculated). Then you would be able to state that of the X number of students admitted to both schools, Y chose to attend school number 1 and Z chose to attend school number 2.</p>

<p>Help? Can anyone (Byerly, I know that you've frequently referred to these numbers) provide a link to these kinds of cross-admit statistics? I'm assuming that some of you know where to find them since there is so much discussion regarding them.</p>

<p>rational - hehe youve brought up the point that i was sayaing a long time ago but no one seemed to care. i was very confused as well and after searching myself, imi pretty sure the pie chart is all the university really knows. its pretty flawed and thus, i dont think using cross-admit data is really worth looking at at all.</p>

<p>many students dont even fill out the little "self answer form" and all the school really knows is where some of these students ended up instead. they dont know much more.</p>

<p>No one can provide a link, because the information is not publicly available. Sometimes people with special access to a particular admissions office -- like those of us who are alumni interviewers and/or former admissions staffers -- can sometimes get this information directly from admissions officers in person. So it's quite possible that a poster has accurate info, even if s/he does not provide a link. Your trust in the numbers depends on your trust in the poster. I have not seen anyone on this board impersonating a Harvard alumni interviewer, if that's any comfort.</p>

<p>You're right about the value of that Stanford.edu pie chart. It isn't as valuable as one thinks.</p>

<p>The revealed preference rankings, however, can give us a sense of the schools cross-admit performance relative to each other.</p>

<p>But even the revealed preference rankings are very very flawed. Keep in mind that the people suurveyed arent necessarily even thinking about the top schools or applying there. Also, I think a 3000 student survey is rather lacking. If the revealed preference rankings were expanded to ask more students it would simply only show a ranking of school's prestige in the minds of those particular students.</p>

<p>
[quote]
Based on information obtained from students admitted to the University who chose not to attend, Rapelye confirmed that most of Princeton's cross-admits choose Harvard, Yale, Stanford or MIT over Old Nassau.</p>

<pre><code>Losing students to Harvard is nothing new, said Chris Avery, a professor at Harvard's Kennedy School of Government who researches the college admissions process.

In a 2000 study of 3,000 highly-qualified students admitted to the nation's most selective colleges, Avery found that in head-to-head battles, schools almost invariably lost out to Harvard.

[/quote]

</code></pre>

<p><a href="http://www.dailyprincetonian.com/archives/2004/10/07/news/10999.shtml%5B/url%5D"&gt;http://www.dailyprincetonian.com/archives/2004/10/07/news/10999.shtml&lt;/a>
<a href="http://post.economics.harvard.edu/faculty/hoxby/papers/revealedprefranking.pdf%5B/url%5D"&gt;http://post.economics.harvard.edu/faculty/hoxby/papers/revealedprefranking.pdf&lt;/a&gt;&lt;/p>

<p>Yes, as Marite correctly points out, the revealed preferences study was a study of students who HAD applied and had been admitted to more than one desirable school. Those were presumably students who were highly motivated to read the viewbooks carefully, talk to high school classmates or neighbors who had been admitted to one or another of those schools, and to visit each campus if at all affordable. Back in 2000, those students couldn't have visited CC ;) but they had an important decision to make, with more than one offer of admission in hand, and they presumably gathered what information they could to make that decision. </p>

<p>When people have a choice between Harvard and anywhere else, they usually choose Harvard. That's no surprise to anyone (like me) who has visited Harvard. (I have traveled to most of the more famous universities in the United States on business trips over the past two decades.) The whole issue of "yield" is why all colleges admit more students than their entering class can hold--but Harvard has to admit the fewest "extra" students, because it can count on most students it admits enrolling for the upcoming school year. A lot of colleges admit more than twice as many applicants as will eventually enroll, because the stronger applicants admitted by those colleges enroll somewhere more desirable--like Harvard, for instance.</p>

<p>oh cuz on the revealed rankings i remember reading in the overview or watever that it was 3000 students who attended schools that are considered "ivy feeders" and have many top school acceptances each year. maybe i was wrong =P but its still not too accurate. i dont think anything can accurately show where more people will go. haha but i think most people will agree that harvard is first. after that its more of a toss up for the next couple of schools.</p>

<p>Hi, Shrek, you are remembering a different sampling, published by the Wall Street Journal, that looked at certain colleges (not as broad a sample as the revealed preferences study) for information about placement at postgraduate professional schools (as I recall the point of that study). That study was done much less thoroughly than the revealed preferences study, which so far is the gold standard of cross-admit research, being based on a good data set to which careful analysis was applied. </p>

<p>Follow the links Marite kindly provided for more information on the revealed preferences study. The second link appears to be the complete text of the study itself, for your downloading and reading pleasure.</p>

<p>In terms of cross-admit "victories" Harvard is the clear winner.</p>

<p>That doesn't mean it's a better experience or a better education.</p>

<p>Actually, the Wall Street Journal graduate school placement rankings, the study you mention, tokenadult, is done quite thoroughly despite a flawed methodology. It is also highly informative. </p>

<p>The editors determined the top five grad schools in law, business and med, and obtained data from those schools about the undergraduate affiliations of the students attending those top grad schools, and then created percentages of students from Harvard attending those top 15 program. Harvard did come in first, with 21.49%. </p>

<p>However, the main flaw in the study was their choice of "top graduate schools"--they excluded Stanford Law and Stanford Business, putting Stanford at a massive disadvantage. Some argue that the "East Coast bias" of the editors prevented Stanford from taking a top-three spot, which are held by HYP in that order. </p>

<p>It used data like this: <a href="http://www.law.harvard.edu/admissions/jd/colleges.php%5B/url%5D"&gt;http://www.law.harvard.edu/admissions/jd/colleges.php&lt;/a&gt;&lt;/p>

<p>Hanna and Marite:</p>

<p>Thank you for your comments. Hanna, you must be correct in saying that the actual cross-admit data are not available to the general public. I certainly can't find them anywhere and no one has stepped forward with any verifiable source (though plenty seem to claim they have one!)</p>

<p>To both of you, I followed your implied suggestions and contacted a couple of alumni I know from Yale and Princeton who had both (in the past) been involved in volunteer admissions work. Both quickly pointed out that their information might not be current and that they had mostly anecdotal information to provide. Both, however, did their work in large urban areas in the northeast and, interestingly, they seemed to agree on the following points. First, both said that in their experience, each of their schools lost more students to Harvard than they won in the cross-admit wars. As to the contest between their two schools, they actually seemed to agree that it was about a draw, with each school doing better with certain types of students. Princeton tended to win the battle with Yale when it came to students who leaned toward the hard sciences and engineering. Yale tended to take more of those interested in the arts and literature. There were many exceptions but that was the general tendency. As to MIT and each of their schools, the greatest overlap seemed to be with Princeton rather than with Yale, since most were science/engineering types. MIT seemed to have a slight edge over Princeton among these students though it was not overwhelming particularly in certain disciplines such as math and physics students. There was less overlap with Yale, but where it existed, MIT was usually the winner (again, presumably because of the type of student in question). </p>

<p>In the battle with Stanford, both said that at the time they were involved there was a definite edge for both Yale and Princeton over Stanford though they understood that this was probably due in large part to the fact that they were interviewing in the northeast. Both seemed to feel that since Stanford had a real recruiting advantage on the west coast and particularly in California, that, nationally, it was roughly a draw. </p>

<p>Finally, the Princeton alum pointed out that the statement Marite quoted from the Daily Princetonian article "Rapelye confirmed that most of Princeton's cross-admits choose Harvard, Yale, Stanford or MIT over Old Nassau," was a misunderstanding by the reporter and that what was meant was simply that, of the students choosing not to enroll at Princeton, the majority ended up at one of those schools. It was not meant to suggest that Princeton loses its cross-admit battles with all of them (though I suspect they do when it comes to Harvard). I wasn't able to confirm this interpretation with the Princeton admission office. If this interpretation is correct, then this information is no more interesting than the Stanford pie chart which also says little about the decisions of cross-admits. I suspect that it's the case for all of these schools that the students who turn them down do so primarily to attend one of their competitors.</p>

<p>There you have it. Interesting but admittedly non-scientific responses (and readily acknowledged to be so). Still, it sounds about right.</p>

<p>As for the revealed preference ranking, it does, of course, examine exactly this issue but, as far as I can tell, the raw data on which the study is based has not been made available. </p>

<p>In the end, I suppose it should be said (as Avery, the author of the preference ranking stated) that these cross-admit studies are not an indication of the quality of education being offered at various institutions. In Avery's words, (as quoted in the article to which Marite provided the link) "basically what that's showing [i.e. the fact that Harvard wins most of the cross-admit battles] is that Harvard's a popular college. It doesn't mean it's better, just that it's popular." (How's that for a little jab on the Harvard board? [smile here])</p>

<p>Thanks to all for your help. I think I've answered my question here.</p>

<p>Rational, thats what the statement means. Of the students who chose not to go to Princeton, most of them either went to Harvard, Yale, MIT, or Stanford</p>

<p>Rational:</p>

<p>Just a slight correction: I made NO comment. I merely excerpted comments from Dean Rapelye of Princeton and provided a link to the only study there is on cross-admits.
This study and the data it is based on are several years old. Not much has changed since then; however, there is one that is worth mentioning: Princeton decided to go head to head with its peers in admissions, and stop the practice, reported in the Revealed Preference Rankings study, of depressing the rate of admissions for students at the 94-98% percentile (which the study claims was pursued to increase its yield). This led to the drop in yield which Dean Rapelye is explaining in her comments.
It should be reiterated that the revealed preference rankings do not reveal anything about the quality of education at one institution or another but focus on the preferences of students who have been admitted at more than one school. Many of the factors that go into their decision are non-academic (location, social scene) and highly subjective.
The study should not be used by students as a guide to selecting schools. It was intended to correct a perceived flaw in the USNWR rankings.</p>

<p>I have been privy to the Harvard cross admit numbers. Let me assure you that ALL elites go to great lengths to calculate accurate cross-admit data ... there is no information that is more useful than knowing how you do, head to head, with your principal "rivals" - unless it is knowing WHY admits chose your school or went elsewhere. (There is a poster - "pinderhughes" - at "the other place" who seems to have access to Brown's cross-admit numbers.)</p>

<p>Schools also go to great lengths to learn the "whys" for marketing purposes. Harvard "loses" about 20% of its admits - or 400 applicants - every year, and it makes a point of learning where they went and, if possible, why. A majority of the losses are to other Ivies and to MIT and Stanford, although many admits who go elsewhere do so to accept large "merit based" or athletic scholarships which Harvard cannot match.</p>

<p>While I cannot reveal precise numbers, I can say that in the past 5 years, Harvard has taken AT LEAST 3/4 of the cross admits from its leading "rivals" as a group (Stanford, Yale, MIT and Princeton.) If anything, the edge is growing. (See the articles about the "winner-take-all" affect in college admissions.)</p>

<p>While the order has not always been the same, more often than not the largest overlap has been with Stanford, the toughest rival (in terms of % of cross admits) has been MIT, and the widest cross-admit margin has been vs Yale (although it was with Princeton for the Class of 2008.)</p>

<p>There are many articles in which Harvard officials and Yale officials acknowledge the historic persistance of this rough 75% cross-admit edge. I have cited some of these articles of several occasions.</p>

<p>"... The only reason Harvard stays competitive (in football), according to (coach) Westerfield, is that it’s Harvard. Three out of four students who get into Harvard and either Yale or Princeton choose Harvard, and it’s no different with high-scoring athletes. “Typically, if I want a kid, I get him,” Westerfield says." </p>

<p><a href="http://www.thecrimson.com/article.aspx?ref=349217%5B/url%5D"&gt;http://www.thecrimson.com/article.aspx?ref=349217&lt;/a&gt;&lt;/p>

<p>"Harvard’s yield—the percentage of accepted students who choose to enroll at the College—is around 80 percent, which is 10 to 15 points higher than its closest competitors. And according to Leverett Professor of Mathematics Benedict H. Gross, who also serves on the admissions committee, more than three-quarters of students who are accepted at Harvard and one of its top three competitors (Yale, Princeton and Stanford) come to Harvard."</p>

<p><a href="http://www.thecrimson.com/article.aspx?ref=214992%5B/url%5D"&gt;http://www.thecrimson.com/article.aspx?ref=214992&lt;/a&gt;&lt;/p>

<p>[In an inderview. DR=Donald Routh]</p>

<p>DR: Last year, Harvard got 83 percent of students that got into both Yale and Harvard.</p>

<p>ZK: Do you have a breakdown of what kinds of backgrounds those 83 percent come from?</p>

<p>DR: You mean, do we lose the poor students or the rich students?</p>

<p>ZK: Right. Because then you would know exactly how the new policies affect that.</p>

<p>DR: Well, we know we didn't lose any poor students to Princeton, even though they took the loan obligation out of the package for students under a $40,000 income. And actually, Harvard [got] 83 percent with students believing they were going to have $1,000 extra in self-help, because [Harvard] didn't make any changes until September of this year, and then they made it retroactive. That was the strangest decision that any of us had ever heard.</p>

<p>ZK: Perhaps a large part of the 83 percent were students from middle-income families.</p>

<p>DR: I don't think it has anything to do with financial aid.</p>

<p>...</p>

<p>YH: So would you call this a bidding war?</p>

<p>DR: In overlap, a group of schools with a natural order of selection--nobody at Yale likes you to tell them that 83 percent choose Harvard, but it's a fact..."</p>

<p><a href="http://www.yaleherald.com/archive/xxvii/1999.03.04/front.html%5B/url%5D"&gt;http://www.yaleherald.com/archive/xxvii/1999.03.04/front.html&lt;/a&gt;&lt;/p>

<p><a href="http://www.harvardindependent.com/media/paper369/news/2004/10/07/Forum/From-Cambridge.To.Cali-747111.shtml%5B/url%5D"&gt;http://www.harvardindependent.com/media/paper369/news/2004/10/07/Forum/From-Cambridge.To.Cali-747111.shtml&lt;/a&gt;&lt;/p>

<p><a href="http://www.yaleherald.com/archive/xxx/2000.11.16/features/front.html%5B/url%5D"&gt;http://www.yaleherald.com/archive/xxx/2000.11.16/features/front.html&lt;/a&gt;&lt;/p>

<p>Get worried!</p>

<p><a href="http://www.thecrimson.com/article.aspx?ref=100552%5B/url%5D"&gt;http://www.thecrimson.com/article.aspx?ref=100552&lt;/a&gt;&lt;/p>

<p>Get even more worried!</p>