EA v. RD: A statistical analysis

Colleges and Universities A-Z Massachusetts Institute of Technology
1

<p>As an RD admit and EA deferral, I wondered to myself what exactly it was that separates EA admits from the deferral pile? It is straight numbers, or are there outside forces at work here? Is the admissions office telling the truth when they say that perfectly valid applicants are deferred for no reason other than a lack of space? Or are they capricious and evil? I ran SAT I scores from both RD and EA admits as reported on the CC threads to compare. Of course, admits v. rejections will result in a woeful underreporting error. If the student only took the old SAT I, I averaged their two sections and added that to their total.</p>

<p>This is what I found.</p>

<p>EA admits:
n=24
Median SAT score: 2270
Std. dev: 129.2</p>

<p>RD admits:
n=46
Median SAT score: 2280
Std. dev: 132.9</p>

<p>Surprising, then, the EA admissions pool actually has a lower median SAT score. This is probably because the RD pool is more competitive, with thousands more applicants. In any case, there is no statistical evidence whatsoever that either pool is fundamentally stronger than the other (p=.29).</p>

<p>The minimums for the two pools were actually exactly the same, save for two statistical outliers of 1780 in RD and 1890 in EA. Both were URMs; make of that what you will. As you can imagine, the distribution is heavily skewed rightward, with a full 25% of the RD pool scores coming in between 2360 and 2400.</p>

<p>In conclusion, MIT Admissions was wholy consistent in choosing their class based on SATs. The admit pools are almost identical if only SAT I scores are considered. Some outside factors must be at work (charming personalities, crazy EC successes, etc.)</p>

<p>But of course, therein lies the problem. I'd like to compare EA admits vs. EA deferrals, then RD admits, but the latter group is quite underrepresented in the Official RD Decision thread (deferrals didn't write that they were deferrals). That would give us a clue whether or not you have to be a much stronger applicant in order to survive a deferral. Also interesting would be to add up SAT I+3 SATIIs for a composite testing score (4800 max). We actually have the data for that, but I didn' think of it earlier and I can't be arsed to look through the threads again. Someone else can pick that up. And that concludes my little analysis (AP Stats did teach me something!)</p>

<p>Discuss.</p>

2

<p>I don't think one has to be "a much stronger applicant" to get deferred->admitted. I don't have data to back up that claim except for my own anectodal experience; I only have a 2260 SAT and no activities that would make me stand out as truly exceptional compared with the rest, and yet I was still admitted. I think (again, this is only my opinion) that once you reach a certain level of qualification what is going to make them look seriously at your application are the touchy-feely subjective things like interviews and essays. I think it would be interesting to know the percentage of defer->admits that went through the interview.</p>

3

<p>No, man. Just... no.</p>

<p>Application strength isn't proportional to SAT scores. Yeah yeah point beaten to death already but it not only isn't that simple, it just... ISN'T.</p>

<p>And EA admits are capped. Only 30% of the class can be made up of EA kids and generally the yield is higher so only something like <400 can be accepted. So they take kids that they want to lock in immediately, I dunno what exactly that means, maybe they stick out in some way during the EA round, maybe they're really qualified URMs, whatever. And defer the rest, strip them of their EA status, and evaluate them in context of now 1000+ spots and RD applicants with no consideration for whether or not they had applied early.</p>

<p>I don't think either is stronger (deferred->admit or EA admit), but by percentages you'd have to be luckier to get in EA.</p>

4

<p>Exactly, that's why I wrote this all up... you can say all you want that app strength isn't proportional to SAT scores, but it's meaningless without solid numbers to back it up. Here are the numbers.</p>

5

<p>I had a friend, an old/good friend, and we basically did a lot of the same things, our stats and numbers and ECs were very similar- i've done programming with this kid and web design etc, we think very similar.</p>

<p>Anyway, I would just like to reinforce the hypothesis (already confimred in large part by the mit.edu blods) that it is the "touchy feely" and subjective things that do distinguish applicants, but here's the catch- there really is no way to game the system or write the "best" essay, because even in apps that are very similar on all counts (like my friend's and mine were) <em>will</em> have some slight distinctions (i might have found one thing that seperated us based on essays, but it's <em>extremely</em> subtle, and only a wild guess, as I havent had any other essays and admitted/deferred numbres to reinforce my hypothesis).</p>

<p>Anyway, this is how I basically told every teacher at my school, many of whom were expecting me to at the worst be waitlisted, and probably get in to MIT, what happened:</p>

<p>"You do the best you can and present ourself as honestly as possible in your essays, feed them into this "black box" we call the admissions office, and hope that you can respect yourself regardless of the result you receive."</p>

<p>That's basically my take on the whole issue i guess =)</p>

6

<p>I partly agree with your assertion, in that SAT scores are (to some extent) taken into consideration in the analysis. However, as a recently admitted student with a low SAT score (1900, 1920), I would like to reinforce the point pebbles has brought up. SAT scores, while definitely important, never make or break a decision. I, too, couldn't get over this until I had actually seen it for myself with my acceptance. They are looking for some "hook" that distinguishes you from the rest of the applicants--some passion or strong pursuit of your interests in almost any field (be it science, music, artist, ect.). SAT scores NEVER serve as this hook (at least in my opinion). </p>

<p>If you think about it, there must be thousands of MIT applicants who fit the SAT ranges you've provided, but only a percentage of them (I doubt they release the actual stats) were admitted. You can confirm this if you analyze the SAT scores of the rejected and deferred students in the EA and RD pools. I'm sure you'll find that the SAT scores of the accepts and rejects don't significantly differ (this is in part due to the skewed distribution of SAT scores of CC members). </p>

<p>Anyway, just my two cents. </p>

<p>BTW, I agree with your final point about the extremely subjective nature of the application. Essays and recs can definitely sway the application.</p>

7

<p>Here it is plain and simple</p>

<p>There is an average SAT score range, and if your within a reasonable margin down (like 2000+) you have enough quantitative qualification. Now they look at what is TRULY important to them: personal characteristics and courseload. This is where interview and essays make the difference. I honestly do not believe there is much of a difference, if any, between a 2100 and a 2400 in terms of capability to succeed.</p>

<p>MIT understands that many people are extremely intelligent, but do badly on standardized tests. They also realize that a low score can mean your just plain dumb. As long are you have reasonable scores they will look into your application further, and the rest is what gets you in.</p>

<p>This reasoning would explain why a high score does not guarentee acceptance and if you look at the reject pool you will see many have near perfect stats. Why? Many, I assume, came off as grade conscious (i.e. retaking a 2300+ SAT, which is great already), and that is never good. The others that had near perfect stats and were rejected either just weren't appealing enough to MIT (bad essays possibly), messed up terribly on the interview, didn't demonstrate passion, or appeared to do things simply to get into MIT, which is the worst thing one can do; for example, volunteering at a hospital even though one doesn't give a damn for the sick and just wants to look good on an app--note that these are all assumptions since I do not know many of the applicants.</p>

8

<p>considering that MIT completely disregarded the writing section this year, maybe you should modify your study to only include M+V.</p>

9

<p>I would also split up the math and the verbal (since verbal isn't as heavily considered as math).</p>