Stats from the Results Data

[number_cruncher]

I've been poking around with the results data to see what kind of statistics can be generated from the available data. So far there's not much, bgk's estimate of 1/10 was a little optimistic. However, here's some data to get the controversy started. From all acceptance results with gpa data and schools with more than 5 acceptances, here's the average GPA of those accepted and number of data points used to compute the average:

1. Stanford University 3.874211 19
2. Cornell University 3.868182 11
3. Carnegie Mellon University (CMU) 3.848182 11
4. University Of California, Berkeley (UCB) 3.835000 14
5. University Of Chicago 3.816364 11
6. Northwestern University 3.795000 10
7. California Institute Of Technology (Caltech) 3.773333 6
8. University Of California, Santa Barbara (UCSB) 3.768571 7
9. Harvard University 3.765556 9
10. University Of Pennsylvania (UPenn) 3.756250 8

Stanford most selective!??! Ahem!

[urbanchic]

And I'm assuming their average GRE scores are over 1550 combined?

[number_cruncher]

not everyone is entering their GRE scores, so the list changes a bit

1. Stanford University 3.873333 1422.2222 18
2. Cornell University 3.855000 1416.0000 10
3. University Of Chicago 3.816364 1434.5455 11
4. Carnegie Mellon University (CMU) 3.814444 1384.4444 9
5. University Of California, Berkeley (UCB) 3.812500 1417.5000 12
6. Northwestern University 3.774444 1373.3333 9
7. University Of California, Santa Barbara (UCSB) 3.772500 1411.2500 8
8. Harvard University 3.765556 1451.1111 9
9. University Of Pennsylvania (UPenn) 3.756250 1431.2500 8
10. University Of Minnesota, Twin Cities 3.700000 1441.2500 8

[repatriate]

bgk's estimate of 1/10 was a little optimistic.

I would be more comfortable giving my numbers if I knew that they wouldn't be displayed with my post. (The current submission page states that they may be.) I am all for providing the database with as much data as possible, but my number combination is pretty unique (one distinctively low and one distinctively high) and therefore would reduce the anonymity of my post if displayed with it.

[number_cruncher]

the data I got only has school, program, GPA and GRE scores. Not sure if there's more in the full DB, but because you don't need to log in to post results I doubt there's any way of tracking individuals. I'll check with bgk.

[repatriate]

the data I got only has school, program, GPA and GRE scores. Not sure if there's more in the full DB, but because you don't need to log in to post results I doubt there's any way of tracking individuals. I'll check with bgk.

I'd more concerned about people who know my numbers IRL viewing the results page than about forumites. Pretty much everyone who knows that info (friends, fellow applicants, people involved in admissions at these schools, supervisors) also knows where I applied. Probably paranoid, but eh.

Edited February 6, 2010 by repatriate

[bgk]

the data I got only has school, program, GPA and GRE scores. Not sure if there's more in the full DB, but because you don't need to log in to post results I doubt there's any way of tracking individuals. I'll check with bgk.

IP address and forum user (if known) are logged too. But these data will NEVER be shared.

I'd more concerned about people who know my numbers IRL viewing the results page than about forumites. Probably paranoid, but eh.

Interesting. To clarify I understood, you'd be happy to submit your scores, etc. if the data were used in aggregate results (i.e. the type above) only?

[repatriate]

Interesting. To clarify I understood, you'd be happy to submit your scores, etc. if the data were used in aggregate results only?

Yes. The scores specifically. That an American applicant was admitted to the PhD program in International Bar Hopping at Stubbleville University is less revealing of who that person is than that an American applicant with a 2.9 GPA and 1590 GRE was admitted....

[origin415]

If there was a way to choose to have your data hidden from the results page and only for statistics, this may help people like repatriate.

[johndiligent]

One factor to consider is that these are nearly all early acceptances, i.e. stand-out candidates that the school is making first round offers to. I would expect to see the GPA average go down in all cases by the end of May.

[noojens]

Edit: nm, answered my question.

Edited February 7, 2010 by noojens