Of course they are different. Correlation is a particular type of dependence. Correlation implies dependence, not the other way around. Hence, they're different, even though one is a type of the other. Learn basic logic before learning probability, please. A little knowledge is a dangerous thing and all that. |
before, whatever the Princeton admit rate was After, some larger amount as the odds of getting into P conditional on having already gotten in to M or H are greater than just the odds of getting into P. Getting into universities are correlated events and therefore dependent. |
No you are completely wrong. This topic is about probability and dependent and independent events are very specific. Correlation has nothing to do with it. In the future I recommend you not insult others when you are the wrong one. |
|
Nash is rolling in his grave. Admissions is not game theory nor is it all about academic stats.
If kid is a great flautist, and academically qualified, and the school band at one of the 10 schools needs someone who plays the flute then they fill an empty bucket and get the admit. In fact that is the best way to look at it. Schools have needs to build their desired community every year. AOs know those needs and go about filling those needs with the “best” available candidate. Applicants don’t know the schools needs in any given year, so there is some randomness in process. |
before, whatever the Princeton admit rate was After, some larger amount as the odds of getting into P conditional on having already gotten in to M or H are greater than just the odds of getting into P. Getting into universities are correlated events and therefore dependent. At the time of application you haven't already got into P or M or H. You don't know any of the results. These are independent events. |
|
I think the OP makes more sense than some claim, at least today compared to the past. When I applied to colleges it was clearly very driven by GPAs and SAT scores; while you couldn't predict what colleges you would get into, it was unusual to get accepted to a more selective college and then rejected by less selective schools.
Today, with GPA and SAT inflation, applicants are much, much more similar on the measurables. Which means it comes down to different factors. Sure, some ECs are clearly more attractive, but it also can come down to the intangibles that may appeal to one college but not appeal to others. In those cases where factors other then GPA/SAT scores are the tie breaker, the applications are (more or less) independent events. And so it does kind of make sense to apply to more schools, at least if the application quality can be maintained.
|
Somebody had to say it.
|
It only 4% offered admission, 96% didn't meet criteria to get an offer, that is almost zero mathematically. |
Did you not read this? Was the math too scary? https://www.stat.cmu.edu/~cshalizi/uADA/13/reminders/uncorrelated-vs-independent.pdf |
Ironic given that you have a much more aggressive tone than me despite not understandingnthe relationship between correlation and dependence. Also, I didn't bring correlation into this, I just said they were dependent (which they are), and you (or someone else) said I was speaking about correlation as if the two were mutually exclusive, when in fact one is a special case of the other. (See the CMU handout for details that are actually specific and relevant) |
|
This is silly. The question is the likelihood of at least one acceptance among students who apply to the same 10 schools each with a 4% acceptance rate. The answer will be between 4% and 33% (OP's math). But there's nothing to narrow it down within that range. The upper bound happens when the acceptances distribute randomly across the 10 schools (the schools may as well be pulling names from a hat), the lower bound happens when all choose the same subset of students (the schools are all using the same algorithm). All an individual student can know is their personal odds are not worse if they apply to more schools. But they already knew that.
|
| Simple probability does not begin to touch the algorithms involved in modern college admissions via the enrollment management industry. Probability is just not how it works. |