Wednesday, 01 August 2007

  • Nerdy Goodness and Awkward Hellos.

    A horde of small Korean children has moved in down the hall from me, and insist on saying hi every time they walk by my door. I always hear them chattering, so I go and look out into the hall to see what's up and am greeted with this trail of high pitched "Hi!"s from 12 inches below me. I'm half amused / half annoyed by it, really.

    Anyways, I have been far too busy doing lab work and socializing and doing nothing that I simply neglected blogging (which I seem to do a lot, judging from my blogging frequency... the fact that the ACTs, which were in March, are still on my front page probably isn't a good sign). However, lab recently took a turn for the worse when
        1. my cells became contaminated and
        2. some of Tejas' (the grad student) experiments suggested that the cells had mutated and
        3. the machine I use to measure my transfections broke and
        4. the professor who owns said $15,000 machine happens to be on vacation until next week.

    This all happened between the hours of 3pm Monday and 5pm Tuesday. The week got off to a splendid start, let me tell you. I have very little usable data, and I'm giving my first presentation this Friday. Problem? I thinketh so.

    So, instead of doing actual lab work, I've been screwing around with my Rubik's Cube and staring blankly at college applications. UChicago wins the best essay prompt by far with its short skit option ("And yes I said yes I will Yes"), MIT wins for cool-ness of the online application homepage, and Princeton wins for best looking college application in general (this is more a matter of personal design preference, though). UMich wins the award for failing to put the online application up when expected.

    Anyways, so today was my second day in a row of not doing anything in lab except waiting for further results on the extent of the mutation and for new cells to grow up. So I thought, gee, why not make a spreadsheet charting the chances of me ending up at each college I'm applying to based on acceptance ranks, possible combinations of acceptances, and school of choice in each said combination? Because this, of course, is what most people do when bored in lab and with nothing else to do.

    (Facebook does not work well in lab)

    So anyways, with my recently downloaded Firefox (god bless it) I went on this hunt for college data and tried to figure out my chances at each college. The chances might seem a bit high, but I'm female and my scores put me in the upper range for all the colleges I'm applying to (and I'm a super-legacy at ND and in-state for UMich), and I wanted all my chances of acceptance to be prime numbers.

    Yes, I am a nerd.

    Anyways, in alphabetical order:

    Accept. Reject.
    Harvard 11% 89%
    MIT 29% 71%
    Princeton 11% 89%
    U. Chicago 41% 59%
    U. Michigan 97% 3%
    U. Notre Dame 71% 29%

    So after that, I figured out all the possible combinations of college acceptances I might get, and then found the probability of each one- so say, my chance of being accepted to all 6 would be (.11)(.29)(.11)(.41)(.97)(.71). In other words, 0.1%. Reassuring, right? And then for the chance of being accepted, say, to 4 schools, it would be the admittance rates of those four times the rejection rate of the other two.

    Make sense? Good.

    So after all this, I then drew up a list of my current college preferences (for instance, Harvard>UMich), and used that to determine which school I would end up at in each combination. From that, I took the sum of each instance in which I would chose a certain college and that's my chance of ending up at that school. Cool, eh? I though so, too. I actually completely forgot to go to lunch because I was so into this.

    It was really interesting, though. I mean, there's only 2 instances where I would go to ND- if I were rejected everywhere but and if I only got into it and U of M. However, the latter is statically extremely likely just because I have such a good chance at both of them. Happily, I only stand a 0.29% chance of living in my parents' basement the rest of my life.

    Anyways, because I am not only stats-a-holic but also a scientist who likes graphs, I decided to make a scatter plot, as demonstrated in Figure 1 below. It was really interesting to see how the stats worked out- I mean, U of M and ND pull up the rear of my list (ND simply because I just like the other schools- besides Michigan- better, not because I still wouldn't go there in a heart beat; UMich because my opinion really and truly places it in the way, way back), and yet they're number 4 and 3, respectively. Really, in order to figure out the rank of the other colleges on my list you'd have to repeat my entire shpeal, and if you actually do so you need a hobby. Desperately.


    (this is Figure 1)

    And this is the most productive I've been in the last 48 hours. Oh, and HSHSP is awesome, but I'll blog all about that later.

    Off to make the Koreans feel uncomfortable.

    Currently Listening
    White and Nerdy
    By Weird Al Yankovic
    see related

Comments (2)

  • pbaranay
    Heh, college stats. For a while I had this idea of making a big spreadsheet listing the admission percentage and median SATs for all the schools I was applying to...and then I realized that most schools wanted to apply to had somewhere between a 10-15% admissions rate, so I was basically screwed no matter what happened.

    Or so I thought. Apparently the Admissions Gods had other ideas. Now I'm a proud member of the MIT Class of 2011. Hooray!

    Oh, random quibble: where'd you get the 29% admissions rate for MIT? I wish it were that high, but I'm pretty sure it's more like 15% if you consider only domestic applicants.

    Damn, I think I'm turning into a blog stalker. That is not good.
  • jzrocker
    omfg, you can't get any nerdier. omggggggg, don't infect me because this is sick-you have a graph and everything gahh
  • Choose Identity

  • Give eProps (?)

  • New! You can now edit your comments for 15 minutes after submitting.

About this Entry

Who recommended?

Who gave the eProps?

2 eProps from:

1 eProp from: