Top Law Schools

elterrible78 wrote:This is similar to a thread I started a couple days ago regarding the relative importance of numbers among T14 schools. In the analysis there, I only included the 2011-2012 application cycle, and based my charts on models that did not consider URM status. There's some instructive stuff there, I think, particularly regarding YP and splitter and reverse-splitter friendly schools.

This time around I am including data from the past two application cycles, and built URM status into the model. For some reason I can't get my software to handle more than two app cycles, so I'll see about working that out. In any case, below you will find both a table with the relevant numbers for each school, as well as rank-ordering of the top-14 schools based on the importance of numbers to their decisions, the bump one gets for LSAT and GPA increases, as well as the bump at each school for applicants with URM status.

Here is the table with data for each school:



The "R2" is the percentage of the variation on admissions decisions (whether you are admitted or rejected) that can be explained by the model used. The first column is the LSAT/GPA model only, and the second is the URM/LSAT/GPA model. The rest of the columns are the log-odds increases for a one-point increase in the measure (for URM, it's just the increase in the likelihood of admission for an applicant with URM status, GPA and LSAT scores being equal).

The next table is just the rank-ordering of the schools for the different measures:

Suralin wrote:

elterrible78 wrote:This is similar to a thread I started a couple days ago regarding the relative importance of numbers among T14 schools. In the analysis there, I only included the 2011-2012 application cycle, and based my charts on models that did not consider URM status. There's some instructive stuff there, I think, particularly regarding YP and splitter and reverse-splitter friendly schools.

This time around I am including data from the past two application cycles, and built URM status into the model. For some reason I can't get my software to handle more than two app cycles, so I'll see about working that out. In any case, below you will find both a table with the relevant numbers for each school, as well as rank-ordering of the top-14 schools based on the importance of numbers to their decisions, the bump one gets for LSAT and GPA increases, as well as the bump at each school for applicants with URM status.

Here is the table with data for each school:



The "R2" is the percentage of the variation on admissions decisions (whether you are admitted or rejected) that can be explained by the model used. The first column is the LSAT/GPA model only, and the second is the URM/LSAT/GPA model. The rest of the columns are the log-odds increases for a one-point increase in the measure (for URM, it's just the increase in the likelihood of admission for an applicant with URM status, GPA and LSAT scores being equal).

The next table is just the rank-ordering of the schools for the different measures:

Loved that other thread you made, love this even more. Thanks for doing this!

Suralin wrote:This probably isn't a valid way to look at splitter friendliness, but NYU, Georgetown, and Northwestern have the largest rank differences (5, 7, 7) between LSAT and GPA for traditional splitters--e.g., NYU is 1st for how much a higher LSAT increases the likelihood for acceptance and 6th for how much a higher GPA increases that likelihood. The same schools also rank highly (2, 4, 5) for how numbers-based they are.

ETA: I guess UVA's famed splitter friendliness only applies to ED applications?

Suralin wrote:This probably isn't a valid way to look at splitter friendliness, but NYU, Georgetown, and Northwestern have the largest rank differences (5, 7, 7) between LSAT and GPA for traditional splitters--e.g., NYU is 1st for how much a higher LSAT increases the likelihood for acceptance and 6th for how much a higher GPA increases that likelihood. The same schools also rank highly (2, 4, 5) for how numbers-based they are.

ETA: I guess UVA's famed splitter friendliness only applies to ED applications?

elterrible78 wrote:

Suralin wrote:This probably isn't a valid way to look at splitter friendliness, but NYU, Georgetown, and Northwestern have the largest rank differences (5, 7, 7) between LSAT and GPA for traditional splitters--e.g., NYU is 1st for how much a higher LSAT increases the likelihood for acceptance and 6th for how much a higher GPA increases that likelihood. The same schools also rank highly (2, 4, 5) for how numbers-based they are.

ETA: I guess UVA's famed splitter friendliness only applies to ED applications?

Suralin, did you see the splitter/reverse-splitter friendly graphic in the other thread? And as far as UVA's splitter friendliness and ED...I have a variable for ED, but I am just not all that confident that people report it accurately on LSN. It might be worth taking a look, though...

And yeah, it'd be interesting to see exactly how the schools rank for ED boost.

elterrible78 wrote:

And yeah, it'd be interesting to see exactly how the schools rank for ED boost.

I'll take a look at that in a bit, but for now, just in general, EDing at UVA made you 1.6x more likely to get in for the past two cycles. I'll take a look at it in more splitter-specific terms here, shortly.

ETA: That increased to 2.7x more likely for splitters. While I don't know how UVA stacks up against other schools for being splitter-friendly with its ED, it splitters seem to do better with it there than the average applicant.