Jump to content

Recommended Posts

Posted

I'm afraid you have your math backwards, as .15^9 would be the odds of being accepted by all nine universities, not rejected. You have to take the rejection rate (.85 in the above example) as the basis for calculating the probability. Given a rejection rate of .85, applying to nine schools leaves a 23 percent chance of an applicant being shut out, ceteris paribus. Applying to 14 schools reduces the likelihood of a shutout to 10 percent,19 schools to lower it to less than five percent and 29 schools to make it less than one percent.

Thank you both. So I have only about a one in four chance of not ever having a PhD. That doesn't sound so bad. Also, I don't think there are 29 schools that I like. I think I'll stick with the nine I applied to and hope for the best. So far I haven't heard anything from any of the schools, but I checked online and I know that all the apps are complete.

Good luck to me! (And to you!)

Posted

The main implication this has is that during the process, once you have acceptances/rejections, you should update your probability of admission accordingly based on new information about G. You can't consider the events independent, given imperfect information about G. For example, if I'm rejected by Rochester, the probability of being accepted by HYP is reduced substantially, because new information changes my G-distribution. If you were close to certain that your G was at a certain level, you could probably treat the events as independent, and calculate probabilities the way you've been doing (i.e. 1-p^n). As it is, even ex ante, I think you probably have to consider your G-distribution when considering the number of schools you want to apply to, and how many in each tier you want to apply to. For example, if I thought my G distribution included a 1/5 chance of being an excellent applicant, 3/5 chance of being a mid-tier applicant, and a 1/5 chance of being a low tier applicant, and if I thought that being an excellent applicant meant I would have a 100% chance of being accepted to any given university I applied to (assuming for the sake of argument, all universities were of the same tier), being a mid-tier applicant meant I had a 20% chance of being accepted, and being a low-tier applicant meant that I would have a 0% chance of being accepted, then your probability can't be figured out from 1-0.8^n, but has to be (1/5) + (3/5)(1-0.8^n) + 0. This would have implications on the appropriate structure/number of your applications, dependent on your utility/cost function.

As a fun exercise, treat G as one's position among the applicant pool, and account for differences in selectivity. Give your answer as code for an R function. :P

My opinion is that this whole notion of an objective selection process based on perceived quality is wrong. Whatever objective criteria probably get swamped by all the random stuff in the error term: the psychological state of the readers, the carelessness of grad admissions department staff (and/or USPS, ETS etc), whether your file was read at the start or at the end, the current state of intradepartmental politics, blah blah blah. Fate, fortune, chance...

Posted

If you people keep this geeky statistics talk up, I will go tell the members of the adcom in my department to whack you. Fun = good. Geeky stats talk = bad. Geeky R coding = worse.

I know. I'm on my 367th line of R code this evening. Don't you be like me, kiddies.

Posted

I'm afraid you have your math backwards, as .15^9 would be the odds of being accepted by all nine universities, not rejected. You have to take the rejection rate (.85 in the above example) as the basis for calculating the probability. Given a rejection rate of .85, applying to nine schools leaves a 23 percent chance of an applicant being shut out, ceteris paribus. Applying to 14 schools reduces the likelihood of a shutout to 10 percent,19 schools to lower it to less than five percent and 29 schools to make it less than one percent.

Yes, you're correct. The consequences of rushing the post :(

Posted

It's been a while since I've taken math stats/probability theory, so this might be a bit off, but I think it's a bit misleading to consider the probability of acceptance at different universities as independent events. I think the better way to think of it is in terms of conditional independence; your probability of being accepted by different universities is independent conditional on a third random variable, namely the quality of you as an applicant/your application. or Pr(H∩P|G)=Pr(H|G)Pr(P|G), where H represents chances of getting into Harvard, P represents chances of getting into princeton, and G represents the quality of your application.

The main implication this has is that during the process, once you have acceptances/rejections, you should update your probability of admission accordingly based on new information about G. You can't consider the events independent, given imperfect information about G. For example, if I'm rejected by Rochester, the probability of being accepted by HYP is reduced substantially, because new information changes my G-distribution. If you were close to certain that your G was at a certain level, you could probably treat the events as independent, and calculate probabilities the way you've been doing (i.e. 1-p^n). As it is, even ex ante, I think you probably have to consider your G-distribution when considering the number of schools you want to apply to, and how many in each tier you want to apply to. For example, if I thought my G distribution included a 1/5 chance of being an excellent applicant, 3/5 chance of being a mid-tier applicant, and a 1/5 chance of being a low tier applicant, and if I thought that being an excellent applicant meant I would have a 100% chance of being accepted to any given university I applied to (assuming for the sake of argument, all universities were of the same tier), being a mid-tier applicant meant I had a 20% chance of being accepted, and being a low-tier applicant meant that I would have a 0% chance of being accepted, then your probability can't be figured out from 1-0.8^n, but has to be (1/5) + (3/5)(1-0.8^n) + 0. This would have implications on the appropriate structure/number of your applications, dependent on your utility/cost function.

This is probably just a formalization of what people are already doing intuitively, and may not actually contradict what you said, but insofar as we're calculating explicit probabilities as in the previous posts (even rough ones), it's probably a good idea to keep in mind how it works out formally.

True (and nice job on fleshing out the math of it), but this is only helpful when one is assessing one's chances once the applications are out, as opposed to when one is deciding how many to send.

I'd also caution that the reasons for rejection aren't always reflecting something lacking in G. For instance, I think that I wasn't exactly an ideal fit for Pitt, given that my interests don't precisely reflect what their major foci are. I think it's conceivable that I could get accepted to universities that are on that same level if I am a better fit for them. There's also an issue of year-to-year quirks in their admissions approaches: school X may be taking in fewer political theory people, because perhaps they admitted more than usual last year, etc. a professor may not want to take on anyone new in a given year. So on and so on.

That being said, I think your approach is still valuable. I just wouldn't necessarily get too discouraged--make sure that there aren't reasons other than G accounting for the rejection, if it is possible.

Posted

I'm pretty sure Duke is not done sending out acceptances. I have talked to my POI and been told verbally that I will be accepted but have received nothing official, so either he is messing with me, or they still have a few to go....

Posted

I'm pretty sure Duke is not done sending out acceptances. I have talked to my POI and been told verbally that I will be accepted but have received nothing official, so either he is messing with me, or they still have a few to go....

Good to know. Also what does POI stand for? "Professor of Interest" is the only thing I can come up with.

Posted

I'm not above whacking my quasi-countrymen. It's the Italian-American in me.

The first generation Italian-American in me is regretting not making an offer the schools couldn't refuse....<_<

Three of my schools have acceptances up and I've gotten nothing.

Posted

Also in at Wisconsin, notified by letter.

I got an email with a scanned copy of an acceptance letter from UW Madison late last night, which is late afternoon over there. I'm hoping that there will be a weekend at a similar time for other universities so that they can all pitch in to fly me out. Are you going to their weekend?

Posted

I got an email with a scanned copy of an acceptance letter from UW Madison late last night, which is late afternoon over there. I'm hoping that there will be a weekend at a similar time for other universities so that they can all pitch in to fly me out. Are you going to their weekend?

Congratulations! (And to you too, Tufnel!)

Posted

Thanks, man. I still haven't heard anything from UCLA, which makes me think that I'm not going to...

Posted

Thanks, man. I still haven't heard anything from UCLA, which makes me think that I'm not going to...

Yeah, I'm not sure why there's a humongous gap between giving out acceptances and rejections. I mean, I can understand a day, but why would a school need a whole week?

Posted

From what I understand, sometimes it's several weeks. I had a friend who'd already accepted a package at NYU who didn't hear back form Columbia until mid- to late-March. I especially don't understand that for schools who only admit one batch and don't do waitlists...

Posted

Yeah, I'm not sure why there's a humongous gap between giving out acceptances and rejections. I mean, I can understand a day, but why would a school need a whole week?

Right? Considering nothing about the rejection need be tailored, you'd almost imagine they go out first. Duke accepted people via website yesterday, so they probably don't even need to enter email addresses or anything.

Posted

I got an email with a scanned copy of an acceptance letter from UW Madison late last night, which is late afternoon over there. I'm hoping that there will be a weekend at a similar time for other universities so that they can all pitch in to fly me out. Are you going to their weekend?

A quick question, did you inquire them via email or did they just sent that to you ? I am also not in US and not sure about the results.

Thanks.

Posted

Are you going to their weekend?

I'm planning on it. I still need to see how scheduling and such plays itself out.

Posted

I'm planning on it. I still need to see how scheduling and such plays itself out.

Translation: "Man, all these hoes up in my game, wish they'd just back up off a brotha. Let a playa play."

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...

Important Information

This website uses cookies to ensure you get the best experience on our website. See our Privacy Policy and Terms of Use