If You Score a 2300+, You Have a 99% Admit Chance at HYPM

Applying to College
1

<p>This is a continuation of a conversation in the last thread (Ivy League Admissions Difficulty is Exaggerated-- please read my first post in that one first), in which I attempt to show that for the SUPERB student (assuming 2300+ SAT, 3.8+GPA, 750+SATII's), there is virtually a guaranteed chance of admittance at at LEAST ONE IVY LEAGUE school. </p>

<p>However, here is a direct mathematical proof. I know many of you have seen this before, but bear with me this is different.</p>

<p>First refer to this (look at Page 7):
<a href="http://www.economics.harvard.edu/faculty/hoxby/papers/revealedprefranking.pdf%5B/url%5D"&gt;http://www.economics.harvard.edu/faculty/hoxby/papers/revealedprefranking.pdf&lt;/a&gt;&lt;/p>

<p>These graphs plot ACCEPTANCE % vs SAT Percentile Range. From it we can clearly see that for the student in the 99th percentile (we will assume this to be ~2300):</p>

<p>MIT = 50% acceptance rate
Yale = 40% acceptance rate
Princeton = 40% acceptance rate
Harvard = 30% acceptance rate</p>

<p>First two suppositions. First, of course the student that scores high on SAT's likely is qualified in other respects (GPA/reccs etc). But we have taken care of this because this proof is only meant for students with (which we assume will place them solidly in the 99th percentile scorewise):</p>

<p>a). 2300+ SAT's
b). 3.8+ GPA
c). 750+ SAT II's </p>

<p>The second supposition is that the rest of the Ivies will likely admit candidates at a comparable rate with these scores. For the sake of simplicity, we will assume 40% (this doesn't even count the fact that Cornell would probably let everyone with these scores in ;)).</p>

<p>Okay so let's assume these 9 schools, 8 Ivies + MIT, admits someone with SAT scores in the 99th percentile at an average of 40% (once again, look at the data if you dont believe me). </p>

<p>The chances of being rejected at any one school: 60%
The chances of being rejected at EVERY school: (60%)^9 = 1.0078%
The chances of being ACCEPTED AT AT LEAST ONE SCHOOL = 100% - 1.0078% = ~99%.</p>

<hr>

<p>Okay now a lot of you are gonna go "NO WAIT IVE SEEN THIS BEFORE THIS DOESNT WORK BECAUSE SOME PPL HAVE 0% and SOME PPL HAVE MUCH HIGHER %". I agree with this.</p>

<p>But look at our original premise. This statistical calculation only works for students that meets our criteria. So we can base their chances of acceptance not on the cumulative average of every applicant, but based on the admit rates of students with SIMILAR scores in the past. So in essence, we have DETERMINED that these students DO IN FACT have an average chance of 40% of admittance at any one Ivy, and if they in fact apply to all 8 + MIT, they will have a 99% chance that they will obtain at least one acceptance.</p>

<p>Any questions?
What does this show? If you're an overworried overachiever with these stats, relax sit back, and wait for at least 1 Ivy Acceptance (of course unless some teacher decides to screw you over or you've been charged with a felony or something).</p>

<p>Edit: I should win someting for showing this ;)</p>

2

<p>By using scores with people who HAVE gotten in, then you can say that 99% of the people who HAD 2300+, 3.8 GPA, 750 SAT II's WERE admitted.</p>

<p>Not 99% WILL BE admitted.</p>

<p>But obviously, the prediction is that there is a high chance you will be accepted at an Ivy.</p>

<p>Good work.</p>

3

<p>
[quote]
we can clearly see that for the student in the 99th percentile (we will assume this to be ~2300)

[/quote]
</p>

<p>You don't have to assume anything about what the 99th percentile is for SAT scores. You can look it up. </p>

<p><a href="http://www.collegeboard.com/prod_downloads/highered/ra/sat/composite_CR_M_W_percentile_ranks.pdf%5B/url%5D"&gt;http://www.collegeboard.com/prod_downloads/highered/ra/sat/composite_CR_M_W_percentile_ranks.pdf&lt;/a> </p>

<p>The top percentile is actually referred to as the 99+th percentile in a College Board report. (An ACT report calls the highest percentile the 100th percentile.) I also have the revealed preferences working paper open on my computer as I type this. For those of you following along at home, the paper can be found at </p>

<p><a href="http://www.economics.harvard.edu/faculty/hoxby/papers/revealedprefranking.pdf%5B/url%5D"&gt;http://www.economics.harvard.edu/faculty/hoxby/papers/revealedprefranking.pdf&lt;/a> </p>

<p>The first thing to note is that correctly read, the table says the admission probability for a top-percentile SAT score at Harvard was 20 percent, not 30 percent, in the indicated year for the indicated sample. Note too the conclusion of the working paper authors: </p>

<p>


</p>

<p>Lower-tier Ivy League colleges, of course, are more eager to accept students with strong numerical credentials in disregard of other issues than Harvard is. It may be that a student applying with a 2300+ SAT score and reasonably high school grades from a challenging high school has a better than even chance of being accepted to the least-desirable Ivy League college (whichever college that is). But the thread title of this thread promises something for HYPM, and I don't trust the promise. It is very easy for a student with such a high set of test scores, and even with very good grades, to come up short at all four of HYPM. Don't bet your house at getting into one of those four colleges without more than just test scores and high school grades.</p>

4

<p>Token I absolutely agree with that statement. But this is simply meant to show that if we extend the 40% chance to all 8 Ivies on average, and if you apply to all NINE SCHOOLS (+MIT), you stand an EXCELLENT chance of admissions and that indeed was the calculation.</p>

<p>If I had actually just calculated for HYPM it would be:
Chance of rejection at one: 60%
Chance of reject at all four: 60%^4 = 13%
Chance of acceptance to at LEAST one of the 4 = 1-13% = 87%, a much lower number than 99%. </p>

<p>My original analysis extends to all 9 schools assuming they have admit rates similar to those of HYPM for students with those scores.</p>

5

<p>You, and all the onlookers, should really consider the implications of </p>

<p><a href="http://talk.collegeconfidential.com/college-admissions/413821-sat-score-frequencies-freshman-class-sizes.html%5B/url%5D"&gt;http://talk.collegeconfidential.com/college-admissions/413821-sat-score-frequencies-freshman-class-sizes.html&lt;/a> </p>

<p>because not all Ivy League colleges are equal, either in enrollment or in desirability.</p>

6

<p>Okay, this is just getting ridiculous.. there were in 2007 about 5400 people in the u.s. that had 2300+ sats. how many people are admitted for all four schools?</p>

7

<p>There are about 17,000 admits every year at the 8 Ivies + MIT. Total yield in the end is about 60-65% average. </p>

<p>SEVENTEEN THOUSAND ACCEPTANCES GO OUT.</p>

<p>Assuming that 1000 of those have bad SAT II's, GPA's, or Recc's that disqualifies them from the running. Another 500 gets chopped due to the inability to compose a coherent essay (a lot of Fobs). This leaves what, out of the 5400 - 1500 ~ 4000 students? And keep in mind I tightened the criteria. Not only do you need to score in the 99th percentile of the SAT's, I also stipulated that these students should score 750+ on their SAT II's and have 3.8+ on their GPA's.</p>

<p>My guess is that AT BEST, half of the 5400 people scoring above 2300 meet these criteria. Better yet, if you scored a 2350, only about 1800 people in the country managed to do so. The higher your SAT scores are, the better ladies and gentlemen.</p>

8

<p>You can't use random independent events calculations, because acceptance isn't completely random.</p>

<p>If 1 out of 2 people get in, it doesn't necessarily mean that the chance to get in is 50%.</p>

<p>You are assuming that by re rolling the admission process, every time the people who get in (all of them, in the 99% percentile) will be randomly selected. In other words not a single person will have preference. This is obviously not true, since a lot will surely have a spot almost assured.</p>

<p>Even if the differences for the top 1% are very small, you can't ignore them and say it's a completely random process - stressing the word "completely". Therefore you can't calculate the chance of an event happening by multiplying the individual chances - because they are neither independent nor random.</p>

9

<p>To add</p>

<p>for example, if someone gets into harvard, pton, mit, yale and columbia, it is more likely he will also get into the lower schools. Therefore, his chances to get in are not random.
You could argue that someone who gets into harvard will have a 70% chance to get into brown also, which is plausible That kinda messes your assumption that all chances are independent and random.
Or, say someone doesn't get in any of the lower 9 schools. What is the chance he will get in harvard? 99% like you predict? The usual 40%? Or even less? See, it doesn't work</p>

10

<p>Wait hold on. I absolutely stipulated that all these admissions outcomes are independent. I'm not saying there's a 99% chance they get into ALL OF THE IVIES. I'm saying there's a 99% chance they won't be REJECTED BY ALL OF THEM (which translates into his getting into at LEAST one of them) if he equally as an independent 40% chance at all 9 of them. Look at the math, it works out. </p>

<p>Have you taken statistics?</p>

<p>If someone with these stats didn't get into the 9 schools I've listed based on the data I cited, he's defying probability. There's only a 1% chance he can end up being rejected at all these schools. Hold on, let me recruit Molliebat, former MIT blogger and now Harvard PhD program to go over the math.</p>

<p>AND, as a general matter for those of you who are like "stuff like this is impossible to predict or even estimate statistically" --></p>

<p>Do you realize that right now in the world, there are people in Insurance companies calculating the chances you will drop dead of a heart attack tomorrow based on how much you weigh? Or that how much they should charge you extra because of the statistic that RED cars are more likely to crash than silver ones? How do you think Insurance companies assess risk? PROBABILITY. And let's just say if "stats" lied, I'm pretty sure they'd be out of business and not making a profit (which in the long run they would exit the market-- a bit of macroecon ;))</p>

11

<p>I will indeed ask for comments about this pair of threads from AP statistics email list participants.</p>

12

<p>Molliebat (I've PM'd her), former MIT blogger and I think Harvard PhD student now may take a look in the next day or so, so let's wait until someone qualifies makes an objective assessment of this method.</p>

13

<p>dont know why youre wasting your time with this...</p>

<p>its not proving anything</p>

<p>acceptances cannot be determined with some random calculations</p>

14

<p>But it's still pretty fun to read. Who cares if it's completely mathematically sound or not. The underlying message is clear: stop overstressing about it.</p>

15

<p>5400 in US with 2300+ SATs? This is not true.. people superscore.. so that figure should be much higher.. the 5400 is for those who achieve in one sitting</p>

16

<p>
[quote]
let's wait until someone qualifies makes an objective assessment of this method.

[/quote]
</p>

<p>Done. The AP statistics teachers who have replied so far are unanimous in saying that the events can't be taken to be independent in the way that they must be independent to apply the multiplicative rule as you have applied it. Certainly no one can deliver on a promise of 99 percent chance of admission at one of HYPM if the only basis for the promise is a score of 2300+ on the SAT and high grades in high school.</p>

17

<p>Actually the premise was 99% you will not get rejected at NINE schools, not just HYPM.</p>

18

<p>
[quote]
Wait hold on. I absolutely stipulated that all these admissions outcomes are independent

[/quote]

Well that's where it goes wrong. The math correct, but your premises aren't. The events aren't random, so using random-oriented math is simply not the case.</p>

19

<p>I don't think you can simply assume the EC's are taken care of ... I have extremely weak EC's but I am expecting a 2250-2320.</p>

20

<p>PEOPLE CALM DOWN</p>

<p>He is not trying to prove einstein here, he is simply showing that acceptance rates for those who ARE qualified are very good at HYPM and all the rest. </p>

<p>Mabye some of you should rejoice instead of instantly bashing this thread with your pessimisstic moans.</p>