<p>PCP, you can create a statistical model of anything, but if you are missing data - or if there is important data that can’t be quantified – then your model is meaningless. You are just making a fallacious assumption based on statistical modeling. Your data is just an artifact of whatever numbers you put in. In any group of data, you will see patterns emerge, but that doesn’t mean that the patterns you see have a causal relationship to anything else. </p>
<p>One problem you have is simply that you have virtually no data to work with other than anecdotal information. Other than SAT & ACT scores of enrolled students and a rough guestimate of class rank, the top schools don’t provide much in the way of data. They don’t list the number of varsity football players, first chair violinists, or Siemens winners in their class – and they don’t tell you how many they might have turned away. </p>
<p>I’d have to say that CC anecdotal information is probably extremely suspect. For example, I think that kids routinely pad or exaggerate their credentials in “chances” threads; and parents may be prone to change little details in their reporting simply for the sake of privacy. </p>
<p>I actually have had a pretty good ability to predict results when I’ve seen the actual essays submitted or am privy to other inside information --at least when I’ve had the opportunity to learn results. </p>
<p>I’ve probably read and reviewed a few dozen essays from CC’er over the years – there are cases when I was pretty sure that the kid sure to be admitted and I was right – but most of the time I never learned whether the kid was admitted or not, and most of the time I didn’t feel that <em>wow</em> factor that made me so certain that the kid’s application was going to be the key to admission. And even in the wow-factor cases, I couldn’t say with any certainty that it was the essay that did it – kid’s who have good stories to tell together with strong writing skills probably have a lot else going for them as well. </p>
<p>What I do know is that there is a lot of bad information about “stats” and a lot of bad advice on “chances” floating around CC. I think partly it is because the environment tends to discourage students with weaker test scores or more limited EC’s from posting details, and most kids who are waitlisted or rejected probably are in no mood to post a detailed recitation of their stats on line. The vast majority of college applicants and admitted students don’t post on CC in any case. </p>
<p>So it seems to me that you are trying to create a model based on no data whatsoever. At best you’ve got a few anecdotes to rely on – but you don’t have a representative sample of applicants, you don’t have a representative sample of admitted students, and you don’t have anything close to complete list of relevant data points.</p>
<p>I’ve read some essays for kids here, and frankly they have all been what I would consider significantly flawed, to say the least. But I noticed that at least one of those kids rated his essay as “awesome” on a chances thread. :)</p>
<p>On the other hand, I am not an admissions officer, and I don’t think that they are necessarily looking for what <em>I</em> think makes a good essay.</p>
<p>BTW, thanks for clearing up the Gatekeepers candidate confusion. </p>
<p>I too felt that Becca was treated a bit unfairly, and that one take home message was that kids at schools where the GCs had relationships with certain adcoms had a big advantage.</p>
<p>MIT results are out today. Interesting to say the least, but VERY high test scores, rank/GPA were rejected time and time again. All that subjective “stuff” must have had some kind of impact.</p>
<p>Caltech puts a lot more emphasis on stats than a lot of other top schools, including MIT. PCP’s model probably would work better with Caltech than with MIT.</p>
<p>just thought I would copy this from a MIT thread. these numbers are pretty mind blowing.</p>
<p>" 3/14/10-A Day for CC recordbooks
Did some Math here:
Total number of 1600/1600s (what MIT counts) that I know (both in life and on here) who were rejected: 17
Total number of 2400ers rejected: 14
Total number of intel STS semi finalists rejected: 4
Total number of Siemens semis rejected: 8
Total Number of USAMO qualifiers rejected: 5
Admit rate for IMO medalists: 50% (1 of 2 could be skewed)
Admit rate for Vals: 20*%(sorry put 2 before [1/5]!) (I only know a few that applied so this may be skewed)
Admit rate for URMs (both on here and off that I know): 37%
Admit rate for asians (on here and off again): 4.5% (1 girl out of about 22)
Admit rate for caucasians: ~6.74%
Admit rate for females: ~18%
Total number of 2250+ rejected: 48"</p>
<p>This makes me think that the metrics used by PCP to assess “Ivy caliber students” don’t work: 2400 SATs, Intel Semifinalists, USAMO qualifiers, etc… rejected. Obviously something other than top scores was needed for admission.</p>
<p>Indeed. None of the six 2400ers who posted were accepted at MIT. One case is particularly astounding: A student with 2400 on the SAT, 800’s on six Subject Tests, perfect GPA and rank, thirteen 5’s on AP tests (including four self-studies), and numerous and meaningful extra-curriculars was waitlisted from MIT and deferred by Yale Early Action.</p>
<p>In such a case…either the stats are not quite truthful, or maybe the problem is in the student’s recommendations. Could it be a case of the student having all IQ and no EQ?</p>
<p>I think that case is perfectly explainable. Said applicant portrays extreme arrogance and seems to be obsessed with tests (he retook a 790 on the Chem Subject Test and retook a 2370 SAT I). Honestly, I’m not surprised that he didn’t get admitted.</p>
<p>Obviously an obsession for perfection is a reason for rejecting an applicant… I guess you should label me as “arrogant and obsessed with tests” for retaking my 2370.</p>