How much does applying Early Action help?

[Deejaiy]

<p>

Perhaps I’m allowing my day job to bleed too much into my after-hours life, but if a difference isn’t statistically significant, in what sense can it be said to exist?

To draw an analogy that hopefully everyone can understand, if a pre-election poll indicates that candidate A leads candidate B by an amount within the margin of error of the poll, then one can certainly still correctly, truthfully, and factually state that there exists a difference in the poll, and that more poll respondents preferred candidate A, even if the difference is within the margin of error of the poll and thus not statistically significant. This analogy is actually not too far off, since the book’s analysis is based on polling data, as mentioned below.

I’m also unclear why it wouldn’t be possible for a person to judge two rounds of applications with the same criteria, particularly when it’s a (fairly large) group of people judging the applications, and when they judge applications over the course of many admissions cycles. This sort of system seems to work well in other contexts – grant applications to the NIH, for example, are submitted in three groups each year and judged by review panels, and I’ve never heard allegations that any particular round is favored over another.

Grant applications are considered on a rolling basis. EA/RA is not. Each round of NIH evaluations is based on perfect information – by the time the decisions are made, all the applications in that competition cycle have been evaluated. EA applications are judged with imperfect information – at the time that you decide on an early admit, you don’t know what the regular pool will look like, even though early and regular applicants compete for the same slots. These are just some of the many differences. I don’t think NIH grants are a good comparison at all.

we don’t have a preference between the two.

MITChris, you certainly state that you have no preference, but the numbers don’t lie, and the book that I referred to above is very thorough in its analysis. I don’t see anything in your blog post containing the amount of detail (admission rate as a function of SAT score, GPA, etc.) that would be needed to back up your assertion mathematically, and in fact I don’t think the admissions office should be posting such information publicly, since it is too revealing. The analysis in the book was based on data that the authors collected in a survey of college students that they conducted on their own.</p>

<p>I’m not picking on MIT in any way; the only reason I refer to MIT so much is because this is, after all, the MIT forum. Many other schools claim that they have no bias but the numbers indicate very clearly otherwise. These other schools are (IMO) outright lying about their EA/ED preference. For MIT the existence of bias is at least debatable, and I have never claimed that it exists in a statistical sense, only that one cannot know for certain.</p>

<p>In such a discussion it’s easy to miss the forest for the trees. The big picture is that MIT succeeds better than any other school at judging EA/RA applicants equally. This is actually a huge accomplishment, and the fact that there is room for improvement should not detract from your success.</p>