<p>Title is self-explanatory . . .</p>
<p>No, to the extent that it’s humanly possible – applicants are read individually, then discussed in committee individually, then a decision is made on each applicant individually.</p>
<p>Of course at some level applicants are compared. If the admissions committee wants to accept between Y<em>a and Y</em>b # of applicants, and there is a pool of X where X is roughly 10 times larger than Y<em>a or Y</em>b, and if the admissions committee independently processes all X applicants and assigns an “admit”/“waitlist”/“reject” stamp, then unless the admissions committee admits exactly between Y<em>a and Y</em>b number of applicants (which is probably pretty rare), they will either have to shoot people off the accept list, or pick people from the waitlist. They aren’t going to shoot off random people from the accept list, and they aren’t going to pick people randomly from the waitlist, so they are going to have SOME way to automatically rank the applicants roughly into different categories.</p>
<p>Note: this is just me guessing around. It would be interesting to see the most efficient way to admit applicants. One way to figure out how high to set the bar for admission might be to read around 1000 applications from the pool at random blindly (though this is not possible), and then figure out some way to roughly give each applicant a score from 0 to 100. Then students roughly around this bar level would be put into a 2nd round pool, with the 2nd round pool having roughly 3000 applicants (and then reassess to pick out the top 1500 ish).</p>