The meaning of early reslults

[untitled]

I was thinking, if an applicant gets an early result, what does that tell the applicant (if anything)? Thinking about it, if I know that applicant A is accepted at school X, while applicant B is rejected from school X, then I might think that applicant A is in some way "better" than applicant B (granted, there are many reasons this isn't true...).

If applicant B, who was rejected from school X, is accepted to school Y, then I might think, well, applicant A is "better" than applicant B, so applicant A should expect acceptance to school Y.

Now, as I said, there are many, many exceptions. But, at least, and with caveats, one might say that applicant A should see acceptance to school X in this case as a reason to be optimistic about school Y. So, I thought, why not try to represent this in a graph (we are assuming that information from results from the past are relevant this year, as I said, many exceptions).

This humble analysis is still in triple-extreme beta (and the url may change if it gets completed), but might be a way to pass the time. What you see is a circle of school names in bizarro alphabetical order, and contains information from the past two application cycles. Clicking on a school turns it blue, and turns a subset of schools red, green, and, sometimes, brown. Schools that turn green are schools that rejected an applicant which was accepted by the school you selected. Schools that turn red are schools which accepted applicants which were rejected by the school you selected. Brown is the combination of red and green, I think.

So if you select UNC, you will notice that Wisconsin turns red! That is, some applicant was rejected by UNC but accepted to Wisconsin, so if you were accepted to UNC these last few days, that means you should be optimistic about your chances at Wisconsin!(?) Looking at the regular analysis for UNC, we see that the applicant was rejected by UNC's econ program and accepted to Wisconsin's ag-econ. So, ones initial optimism is lessened somewhat.

Anyway, what do you think? Is this at all useful? I wanted to come up with a set of data that would be interesting to represent in a wheel, and this is what I came up with... Any better ideas? It isn't hard and works better just to check each school for rejections, but this pulls that info together into one web-page (and helps me learn flash/flex/flare and representing graphs and trees)...

Also, is this logic at all valid? I know there are a million exceptions, but I'm looking for a reason to be optimistic (and I haven't even heard any news yet), and thought this might be one. Also, we could consider a rejection from school X as reason to be pessimistic about schools which rejected applicants who were accepted to school X (is that the contrapositive?). Is anyone interested in seeing this graph?

[99luftballoons]

We should be careful with the confounding variables. High-theory schools (ie UMinn, Stanford GSB, Princeton) will be looking for different kinds of candidates than more applied schools (AG Econ, b-schools that are not Stanford, etc...) so it is hard to actually rank candidates in a meaningful way on just acceptances/rejects.

Your graph, however, is very pretty and I am going to play with it for hours hoping UMinn rejections come in.

[Sonaar]

Great idea, very useful... but i think the colors should be reversed.

[untitled]

I guess I could switch the colors, but I'm not planning on doing it. I did fix a couple more mistakes in the data that the graph uncovered, and I ordered the graph. The ordering is somewhat arbitrary, so I went with a geographic order, which has a benefit of being fun (and slightly prettier) and not much of a drawback. Let me know if you like the new order (or have suggestions).

[mathemagician]

How do you have the time to do this untitled? No classes this semester/unemployed?

[treblekicker]

Wow, I have a very similar profile with bertthepuppy, except I had more Math and less Grad courses. My university may have been slightly better ranked overall for UG by USNews, but that is pretty irrelevant.

Interestingly enough, the only schools in common for us are UNC and Duke. Makes sense, since I am more interested in Theory and Metrics as opposed to Applied Micro.

Here's hoping I don't follow bert with a Duke rejection. I'd imagine that my letter of recommendation writers have more connections to Duke than Bert does, so I hope that helps!

[untitled]

How do you have the time to do this untitled? No classes this semester/unemployed?

I'm too nervous to reveal anything about myself, but I do have stuff that I should be doing when I've been programing. Luckily, it was especially slow the second half of January, and this made the wait more bearable. Now, after a busy two weeks, it looks like things are getting slow again. Anyone, let me know if you can think of any analyses to do with the tm data.