Lessons Learned

[Modeler]

I really thought I had a handle on the application process because through undergrad and then transferring, my record of acceptances had been impressive; however, with all the rejections I received this year, including from all of my safety schools, it becomes clear that this is a whole 'nother ballgame. With this in mind, I felt I should share the things I learned to help out future applicants. Feel free to contribute your own insights.

This is a little disorganized because right now its my thoughts as I think of them.

What may have been my most dramatic misconception was that US News rank correlated in anyway with selectivity. Based upon US News rank, as well as guidelines published by the universities, this is how I thought my programs shook out (Harvard is excluded from these discussions because Systems Bio (the program I applied there for) is different enough that the lessons may not be the same).

Reach:
Stanford
CMU (COS Program)
Cornell

Solid:
Michigan
Maryland (Scientific Computation)
Penn
Brown

Safety:
Duke (Duke has since raised in the standings so this now looks foolish)
NYU
Northwestern

At some point I link the post with my profile so these make more sense.

With my improved knowledge, this is now how I feel these looked for me (me personally, not generally):

Reach:
Stanford
Cornell
Brown
Penn
Duke

Solid:
CMU (COS Program)
Michigan
Maryland
NYU
Northwestern

The major reason for this reorganization is that in general, the schools that were great matches for my background and interests showed dramatically more interest in me. For comparison, here are my results:

Accepted: Michigan, Maryland
Wait-listed: CMU*, Cornell**
Rejected: Stanford, Penn, Brown, Duke, NYU**, Northwestern***
*Officially wait-listed, still no word, was told wait-list had only three people on it, but the program is small, maybe like 12 new students a year
**Unofficially wait-listed, by which I mean I still haven't heard anything from Cornell, and NYU notified me after the deadline despite others receiving their notification weeks earlier.
***Was offered a spot by a professor in a lab that was unrelated to my interests, told him so and a month later was officially rejected. So this is kind of a win.

Comments on each school:
Stanford - Really chasing a dream here. Too many stellar applicants to seriously be considered
CMU (COS) - I had conducted research in modeling social interactions and I feel this is why they maintained interest despite me not making their posted standards of "we expect all applicants(!) to have a GPA of at least 3.85 [mine was 3.7] and above a 750 on each section of the GRE [Me: V 710 Q 800]"
Cornell - still no word but I mentioned wanting to improve the state of public understanding of CS, and given that Cornell has some reputation for favoring good prospective teachers maybe this is what kept me in the game
Penn - Thought I had a good shot based on historic standards of GRE and GPA. I think what really hurt me here and other private schools is that I come from a state school without a reputation for CS.
Brown - I feel places like Brown, JHU, and others (Notre Dame springs to mind as a more extreme case) are more selective than their ranking would indicate because they have small programs with big brand name recognition. Its almost not worth applying to a place like these without outstanding reasons because on average you won't get good value out of your application fee.
Duke - I could not believe this rejection. I've loved Duke all my life and they accept roughly 30% of American applicants. May have been hurt by my school here as well. Last year their stats indicate only 23 American applicants were not accepted. This stung.
Maryland - they couldn't find my GRE for a long time, but once they found it they accepted me within a week...hmm. I think my background was really good for scientific computation (Numerical Analysis I + II, Dif EQ, Linear Algebra) and it didn't hurt that I got basically 100% in Numerical Analysis and had that professor write a rec for this program. Also I want to point out that Maryland and Michigan, my two outright accepts, responded within the first month. The state schools seemed to really love me. More on this later.
NYU - my base stats were significantly better than their average. I just don't think I was a good match here. Hence the late rejection.
Northwestern - this case shows how little your SOP is read some places. My only offer came from an embedded systems prof, despite no serious interest or background in the area.

And finally, Michigan - This is where I'll be attending. In addition to being a top 15 school, I was offered a big fellowship so this was a no-brainer. I almost didn't apply here based on climate! As detailed elsewhere, I visited after I was accepted. If not for my visit, I would not have the fellowship because I didn't know that I was supposed to request a nomination, or even that it existed. Coincidentally, all the professors I met with are associated with the program that the fellowship is attached to, and they nominated me for it after meeting me and discussing my interests. I was extraordinarily lucky, but it just goes to show how important visits and interviews can be. I also learned from this visit how my application was considered. I was told that some parts of my additional statement were viewed negatively. In undergrad and transfer applications I found success by discussing obstacles I had overcome, specifically having an engagement fall apart and being diagnosed with ADD. For graduate school applications, so long as you have eventually turn it around, it may not be worth mentioning why your early grades were not great. After being told this, I asked why they liked me, and I was told pretty directly that it was my GRE score. This, and the eagerness of Maryland to accept me after they found my GRE, contributes to me believing that generally big public programs may place more weight on your GRE.

If I were to do it over again with the knowledge I have now, I would have attended Penn out of high school, but otherwise taken the subject GRE (may have helped me a lot, just didn't have the time to study) and published papers instead of just talking about potential papers. I also would have look more seriously at big public programs because they have more slots to award and some of them are even better ranked than the private schools I applied to out of brand name recognition.

[MDK]

Wow, very informative Modeler. I would advice every new applicant to read this post before thinking of applying. This would be helpful for us Internationals more than Americans since it would demonstrate how Citizens suffer to secure admission into good graduate school, not to mention Internationals.
And the sentence I liked the most is:

What may have been my most dramatic misconception was that US News rank correlated in anyway with selectivity

[nirosilva]

Modeler,

Great tips, Thanks

[shemz]

Woah nice to know that CMU kept someone with 3.7 GPA on waitlist. I was putting in my high probability list with a CGPA of 3.45; need to reconsider my chances now after going through this thread. Btw I beleive CMU can be cracked based on the fact that I already have a international journal publication in my UG and I am applying for a PhD.

[CalmLogic]

Of course, for computer science PhD admissions, CMU, like Berkeley, is usually mission impossible. As you may have seen, Modeler was applying for their "Computational Analysis of Social and Organizational Systems" PhD program, which surely gets far fewer applicants:

CASOS: Home | CASOS