Does Yale Top Harvard and all others in cross admits?
|repapa|| 05/06/10||Nathan Hale|| 05/11/10||NAFCYN|| 05/11/10||.,,,.,..,..,,.,,,.|| 05/12/10||HLS_Grad|| 05/20/10||.,,,.,..,..,,.,,,.|| 06/17/10||Nathan Hale|| 05/12/10||NAFCYN|| 05/12/10||Nathan Hale|| 05/12/10||NAFCYN|| 05/13/10||NAFCYN|| 05/12/10||Nathan Hale|| 05/12/10||NAFCYN|| 05/12/10||repapa|| 05/13/10||NAFCYN|| 05/13/10||repapa|| 05/14/10||SuperTrader|| 05/14/10||NAFCYN|| 05/14/10||NAFCYN|| 05/14/10||NAFCYN|| 05/21/10||repapa|| 05/22/10||NAFCYN|| 05/22/10||repapa|| 05/22/10||NAFCYN|| 05/23/10||SuperTrader|| 05/13/10||repapa|| 06/03/10||E1Greco|| 06/04/10||NAFCYN|| 06/04/10||E1Greco|| 06/04/10||NAFCYN|| 06/05/10||repapa|| 06/14/10||NAFCYN|| 06/15/10||repapa|| 06/15/10||NAFCYN|| 06/16/10||repapa|| 06/16/10||NAFCYN|| 06/16/10||repapa|| 06/17/10||NAFCYN|| 06/17/10|
Poast new message in this thread
Date: May 6th, 2010 5:58 PM
Subject: The CC thread would certainly suggest so
For three consecutive years Y beats H narrowly and Princeton overwhelmingly in the battle of cross admits. See discussion about cross admits among HYPSM http://talk.collegeconfidential.com/college-search-selection/920080-harvard-yale-princeton-stanford-mit-cross-admits-class-2014-a-17.html
Yale won the CC cross admit battle handily in each of the last 3 years (and more important in the combined 3 years overall). The aggregated three year numbers (more than 100 matchups for each of HY and P) are likely statistically significant. The general picture is that Y easily has the highest win percentage against the other 4 (including about 53-47 vs H and 80-20 vs P),and that the other 4 are about equal in overall win percentage.
Could it be that Harvard suffers from Tufts syndrome and employs the former Princeton strategy of shying away from likely Yale admits to minimize the size of its cross admit pool with the other elites and thereby maintain its high yield? Wouldn't be surprising. You ought to read through the whole thread.
Date: May 12th, 2010 3:03 AM
you're such a fucking loser
Date: June 17th, 2010 6:44 PM
actually idiot i easily beat nycdolt in the prestige race
Date: May 13th, 2010 12:17 PM
The fact that cross admit numbers have apparently changed so little over 10 years demonstrates two verities:
1. The academic pecking order is relatively static and slow to change; and
2. The "revealed preference" ranking is far harder to "game", since even yield rates can be manipulated via the use of early admissions programs, "likely letters", "merit" aid, etc.
Date: May 12th, 2010 11:35 AM
Subject: Here are SOME "hard numbers":
For the Class of 2013, these are the number of admits who wound up matriculating elsewhere ... ie, the "cross-admit losses":
Harvard: 512 (23.5% of admits)
MIT: 604 (36% of admits)
Yale: 651 (33.2% of admits)
Stanford: 732 (30.2% of admits)
Princeton: 889 (40.2% of admits)
For each of these schools, between 60-80% of cross admit losses are to each other.
NOTE: Harvard and Princeton have open admissions; Yale and Stanford fill substantial fractions of the class via "single choice early action; MIT has an open early action program.
Date: May 13th, 2010 5:28 PM
Subject: of course H boosts its yield artificially and controls its
cross admit losses by heavy reliance on its WL - 200+ in 07, 130 or so in 08. That adds a couple of percent to the yield and minimizes cross admit losses. Youve also understated Yales yield by a point or two.
More important, its the absolute number of non-acceptees (i.e. 550 for H vs 630 for Y) rather than the overall yield which is more important in trying to divine the cross admit wins and losses in head to head match ups. The bottom line is there is not any good reliable data to determine what percent of cross admits go where. The CC cross admit data, when aggregated over 3 years, probably approaches reliability and suggests a slight edge for Y over H, though admittedly I would not put much faith in that.
Keep in mind that Y, as some CC poster noted, may lose more to AWS and similar schools than H does, since Y is seen as more of a liberal arts bastion than H.
Date: May 14th, 2010 11:04 AM
Subject: Incidentally ....
The mystery behind the delay in release of the preliminary yield number at Yale is partially solved: they are apparently filling in as many slots as possible from the waitlist before a preliminary announcement - probably trying to beat last year's preliminary yield number.
See the CC thread where waitlistees are currently being called and quizzed as to whether they would "like to come to Yale."
Date: May 21st, 2010 12:05 PM
Subject: Projected yield rate at Yale for Class of 2014
There were 1,940 admitted initially, including a record 732 from the early pool.
If there are zero waitlist admits, and the matriculating class is the same size as last year (1,307), then the yield rate will be 67.3%
If there are 30 WL admits in order to achieve a matriculating class of 1,307, then the projected yield rate will be no higher than 66.3%, down slightly from last year.
Under this scenario, assuming an historically equivalent 88% yield on the early pool admits, 1,326 will have been admitted from the "regular" pool and 663 will have matriculated. for an "open market" yield rate of exactly 50%.
Of course, the projected class size may be higher (originally 1,320 last year before "summer melt") or the number taken from the WL may be higher or lower than my projected number of 30.
Date: May 22nd, 2010 8:11 PM
Subject: You still don't know what "open market yield" is.
In a true "open market" situation, circumstances must be exactly the same with respect to all applicants - ie, they must have immediate opportunities at competitive institutions. This is demonstrably not the case when applicants are invited to apply early only on the condition that they are barred from applying early elsewhere. Indeed the opposite is the case - which is exactly the advantage the SCEA school seeks!
The whole purpose of such a device is to (hopefully) gain a competitive advantage - particularly at the expense of arguably more selective schools.
The scenario is that (1) "School H" attracts the top students-who will only apply there, or will choose it if admitted to both "School H" and "School Y"; "School Y" attracts and keeps the applicants who it would normally "sign" anyway; BUT - (3) "School Y" also attracts and keeps a certain number of early applicants who, having gained admission, no longer see the marginal utility of a "regular" application to "School H" where the odds of such admission are quite low for the "regular" pool. This is true even where - all things being equal - a "School Y" admit in group (3) might originally have opted for "School H" if given a contemporaneous choice.
"School Y" anticipates (and indeed relies on) this inertia in order to snare a few high quality admits who might otherwise have applied to, and been admitted by, "School H".
The success of the gamble is demonstrated by the fact that "School Y" (and "School S") have an SCEA yield rate that is only very slightly lower than the yield rate with the ED device they previously utilized.
This scenario has been analyzed in papers demonstrating (indirectly) why Harvard could afford to drop the early admissions crutch, while Yale could not, and (previously) why Yale was forced to move from ED to SCEA as its number of early applicants dropped relative to the competition.
Very interesting stuff!
See: by Chade, et als (2009)
And, by Avery et al: (2009)
Date: May 22nd, 2010 11:16 PM
Subject: Factually Inaccurate
By the time H abandoned SCEA, Yale had a (much) larger SCEA pool than H and a much lower early admit rate. There has been speculation, including Levin's comments in a Harvard publication, that H moved away from scea because it was losing the early round battle to Yale, and its cross admit advantage had all but disappeared (owing to the fact that Yale had switched from early decision to scea, freeing its early admits to trophy hunt as Harvard early admits had already.
H does not monopolize the best and the brightest. Its gets its fair share, but keep in mind that Yale has been the most selective school in the country (measured by SATs and HS transcripts) in 3 of the last 4 years. Y does not need to play games to compete effectively with H. Keep in mind the data on CC suggesting that in cross admit battles among HYPSM, the big winner is Yale. That data, consistent over 3 years, is at least as good as any other bandied about on these threads.
Date: May 23rd, 2010 11:38 AM
Harvard admits generally don't consider a Yale admission a "trophy"!
But speaking of the little world of CC, did you notice that there are 10 CC Yale admits still hanging on, hoping to get in off the Harvard waitlist?
Of course, maybe they don't REALLY want to get in, but just want a "trophy", eh?
Further, it seems 5 other CC Yalies who sought admission from the Harvard waitlist have already been denied. The 5 Harvard WL admits on the CC site so far came from Columbia, Princeton, Chicago and MIT (2).
16 CCers report having been taken off the Yale WL so far, and the report is that "there will be a second round of WL admissions after May 25" when the number of defectors to WL invitations elsewhere may be clearer.
Date: June 3rd, 2010 10:15 PM
Subject: Actual crossadmit data for h and y from CC
Interesting- over the last 3 years years Yale defeated harvard 28-25, winning 53% of the cross admit battles reported on college confidential. The sample size is probably large enough to be representatve of what's truly happening in the hy competition for cross admits. I would conclude that it's a fallacy that h wins the dominant share of crossadmits from yale.
My conclusion is bolstered by the fact that Yale also fares far better than h in combined winning percentage against the top five-hypsm. That sample pool is approximately 200 on cc. How does h maintain the highest yield when it has serious cross admit problems with y and some of the other super elites? Could it have a case of Tufts syndrome sighing away from the high energy, more dynamic types likely to gain admission to yale, etc
Date: June 4th, 2010 9:11 PM
Author: E1Greco (elite, mysterious, elusive, prestigious)
Hi, NYCFan, or Alex.
My name is Michael, and no, not everything is "borrowed." I have a fulltime roster of 11 staff that are working with me to solidify the community. The registrations so far for just Harvard have exceeded 200 students, staff, and alumni.
I think it's going to be an excellent way to build up the communities within each of the ivies, and other top schools. Being able to directly participate in these cross-discussions will only be a benefit for everyone.
You've long been known to be staunch advocate for dispensing the proper information to the appropriate audience, which is why the Harvard listserv would be the more appropriate medium for this to occur, as the college audience on this site is by-and-large, dead.
Date: June 16th, 2010 11:00 AM
The latest "sample size" of Harvard/Yale cross admits who chose one or the other in the Ewho/Mathacle thread is 24, with 12 opting for Harvard and 12 for Yale. Hardly large enough to generalize, when the total number of H/Y cross admits almost certainly exceeds 400 this year.
Furthermore, I have no idea whether the CC posters are a "typical cross-section" or not, but I suspect NOT. Finally, the tiny sample is not randomly selected, or scientifically selected, but rather SELF selected - which is a no-no when you want to make a projection.
By way of comparison, the "Chances" site utilizes the chess-ranking formula applied 10 years ago by Avery, et al, and can rely on data from 50,000 responders and 250,000 applications.
You might want to read THIS where the developer defends his formula:
Date: June 16th, 2010 5:16 PM
Subject: The CC sample size for HY was 53
over a three year period - i.e. classes of 12, 13 and 14. That would likely constitute a meaningful statistical sample, particularly where the rank order (Y followed closely by H) is the same as is seen in the larger three year tournament of HYPSM in the aggregate (i.e. Yale beat Harvard, and it also beat all others within the group). There were approximately 200 such matchups over the three year period.
BTW, it would suprise me if there were as many as 400 y/h cross admits this year or in any year. Are you aware of any such data?
Finally, the Chances site provides no info on how many direct H/y matchups it has and whether its sample size was smaller than, comparable to or larger than the CC cohort. Its huge confidence limits suggests the number is quite small.