<p>No, he's saying that if they didn't do it holistically, then the adcomms would be unemployed</p>
<p>it's a THEN (correlation) statement, not a BECAUSE (causal) statement</p>
<p>Yes, the way the sentence is structured it's a THEN statement. But how is it relevant at all to anything if its not really causal.</p>
<p>I've heard many different adcomms say it. I think the fact holistic admissions justifies their existence betrays at least part of the motivation for it.</p>
<p>nooooooo</p>
<p>It's just a simple way to allay public fears by making a simple but often overlooked observation, nothing more.</p>
<p>It's a bit fallacious, but not really intended as an argument.</p>
<p>Thanks, 2boysima, for the specific link to a New York Times artlcle </p>
<p><a href="http://www.nytimes.com/2007/04/04/education/04colleges.html%5B/url%5D">http://www.nytimes.com/2007/04/04/education/04colleges.html</a> </p>
<p>that is recent and which refers to Harvard's number of rejected students in one perfect SAT section score category. Note that the journalistic statement doesn't contradict a possible factual situation that Harvard admitted hundreds of perfect scorers on the math section, as indeed its freshman class score profile </p>
<p>College</a> Search - Harvard College - SATĀ®, APĀ®, CLEPĀ® </p>
<p>makes clear it did. </p>
<p>I'm not sure in all cases why other participants on CC mention that they've heard that "Harvard rejected 50% of their perfect SAT scorers." I don't know whether that's even a true statement, although I don't doubt at all the statement that some students with a 2400 on the SAT are rejected each year at Harvard. I opened this thread just to check whether there is any source that would tend to back up that statement. The revealed preferences working paper really doesn't back up the statement, because "perfect scorer" is not the same category as "top percentile scorer." I will also note that the working paper may not provide current information about Harvard's admission practices in the most recent two admission cycles. It surely doesn't provide current information about Princeton's practices (although it may have helped prompt a change in Princeton's practices).</p>
<p>My good friend's son went to H. Perfect Math, perfect SAT2s, but a 780 in the verbal. This was before the three part SAT1s</p>
<p>"I hate this argument. The adcomms are essentially saying that they have to do admissions holistically because otherwise they would be unemployed."</p>
<p>No, they are saying that there are things more important to them than one (or more) three hour "perfomances" which are closely correlated with family income, subject to change as a result of extensive (and expensive) coaching, and which might make for a very dull class indeed.</p>
<p>But if one wants to attend a school where standardized testing is the be-all-and-end-all, why would one want to go to Swarthmore or Princeton?</p>
<p>One of the best researched book on this subject is 1600 Perfect Score: The 7 Secrets of Acing the SAT by Tom Fischgrund. </p>
<p>Amazon.com:</a> 1600 Perfect Score: The 7 Secrets of Acing the SAT: Tom Fischgrund: Books</p>
<p>Tom Fischgrund obtained unpredented access to the files of 1600 scorers. Via Amazon reader, you can read pages 51-53 that describes the percentages of addmission, rejection, and attendance at a number of competitive schools. </p>
<p>For instance, Harvard accepted 53% and Duke ... 100% of all the 1600 applicants. On the other hand only 4% of the Duke 1600 ended up attending Duke.</p>
<p>
[quote]
No, they are saying that there are things more important to them than one (or more) three hour "perfomances" which are closely correlated with family income, subject to change as a result of extensive (and expensive) coaching, and which might make for a very dull class indeed.
[/quote]
</p>
<p>Actually, standardized test scores don't even make it to the list of "very important" factors as listed on the Commond Data Set:</p>
<p>Very Important
Rigor of secondary school record
Class rank
Academic GPA
Application Essay
Recommendation
Character/personal qualities</p>
<p>Important
Standardized test scores
Extracurricular activities</p>
<p>Considered
Interview
Talent/ability
First generation
Alumni/ae relation
Geographical residence
Racial/ethnic status
Volunteer work
Work experience
Level of applicantās interest</p>
<p>anecdotally, it seems that once you cross a threshold "2100" "2250", what have you, at different schools, it doesn't really matter...ie, a 760 = 800...and this is coming from an 800 CR scorer and 800 USH</p>
<p>Of course, I don't believe the Swarthmore list even for a second. If the "very important" factors truly reigned supreme, there would be a much, much broader swathe of SAT scores both within each entering class, and from year to year. And if "racial/ethnic" status are ranked so low, they must find an awful lot of "character" associated with racial and ethnic characteristics. Alternatively, hundreds and hundreds and hundreds of pinky-greys year after year are found wanting in the character department. ;) (And no college dare put "developmental" factors on the list, even when they may easily trump everything else.)</p>
<p>It would be interesting to know if a statistician trained in regression analysis would agree that the most important admission selection factors are the same as those reported by the admission. I suspect not, based on the studies of admission practices that I have read.</p>
<p>The problem is that "a statistician trained in regression analysis" might not know a good application essay if it bit him on the butt.</p>
<p>
[quote]
And if "racial/ethnic" status are ranked so low, they must find an awful lot of "character" associated with racial and ethnic characteristics.
[/quote]
</p>
<p>The Supreme Court decisions in the recent Michigan affirmative action cases really don't allow colleges to admit more than "consideration" of race and ethnicity.</p>
<p>
[quote]
And no college dare put "developmental" factors on the list, even when they may easily trump everything else
[/quote]
</p>
<p>The Common Data Set saves colleges from embarrassing themselves by providing a santizied list of considerations. The colleges merely have to move the choices around to the appropriate columns. Notice that the Common Data Set folk conveniently fail to ask where "recruited athlete" ranks, too.</p>
<p>for what its worth, here is info about princeton -- Princeton</a> University | Admission Statistics</p>
<p>The assertion has been made upthread that admissions offices wouldn't necessarily know their perfect scorer admissions percentages, because this isn't a tracked category.</p>
<p>From the little I have seen of the inner workings of MIT's admissions office, I am confident they have this capability -- they don't track perfect scores for the purposes of admissions, but they do have a database which contains an entry for each student, and which would allow them to calculate the number of x scorers accepted for any score range. I observed them using this database to calculate statistics on individual SAT IIs for their internal records.</p>
<p>"for what its worth, here is info about princeton"</p>
<p>Notice there are no admissions statistics by family income. They have it, of course. But they won't dare post it publicly.</p>
<p>Thanks for post 26. :)</p>
<p>"No, they are saying that there are things more important to them than one (or more) three hour "perfomances" which are closely correlated with family income, subject to change as a result of extensive (and expensive) coaching, and which might make for a very dull class indeed."</p>
<p>Of course, I know what the main argument is. All I'm saying is that adding the statement, "If the policy was different, then you could fire all of us," does not add to the argument.</p>
<p>"It would be interesting to know if a statistician trained in regression analysis would agree that the most important admission selection factors are the same as those reported by the admission. I suspect not, based on the studies of admission practices that I have read."</p>
<p>"The problem is that "a statistician trained in regression analysis" might not know a good application essay if it bit him on the butt."</p>
<p>Right, but that wouldn't be necessary. The statistician trained in regression would simply being doing the regression to determine if the importance of the factors was indeed similar to what is claimed. Since the essays can't be used as numeric input as is the way that GPA and SAT/ACT can be, someone who WOULD know a good application essay if it bit him on the butt would simply rate each essay using whatever scale and the statistician would enter that into the equation as an independent variable just like GPA, SAT, etc. The statistician wouldn't have to be actually assessing the quality of the essay, merely the impact.</p>
<p>
[quote]
Right, but that wouldn't be necessary. The statistician trained in regression would simply being doing the regression to determine if the importance of the factors was indeed similar to what is claimed. Since the essays can't be used as numeric input as is the way that GPA and SAT/ACT can be, someone who WOULD know a good application essay if it bit him on the butt would simply rate each essay using whatever scale and the statistician would enter that into the equation as an independent variable just like GPA, SAT, etc. The statistician wouldn't have to be actually assessing the quality of the essay, merely the impact.
[/quote]
</p>
<p>Try Emory, maybe. Or Duke. Or some other school that is heavily numbers driven in admissions.</p>
<p>Assign a numeric rating to the essays? I'm just chuckling at how conceptually foreign that would be to the admissions process at a school the size of Swarthmore and the way applicants are evaluated. They aren't grading the essays. They are using the essays as a way to get to know living, breathing teenagers and try to visualize each applicant as a part of the campus community.</p>
<p>Doing regression analysis on SAT scores is sending people on a wild goose chase as far as really understanding elite colllege admissions. The thing that to understand is that virtually the entire applicant pool at these colleges has strong stats. It's not that test scores don't matter, it's that essentially everyone who applies has good test scores. The scores of the applicants look pretty much the same as the scores of the enrolled students. So, the scores are largely useless in differentiating a winning application from a losing application.</p>