Top Law Schools

Bad semester dragged down my GPA, coming from UWmadison would receive in state at minnesota looking at Chicago area schools applying early decision to UMN

Threezus wrote:Bad semester dragged down my GPA, coming from UWmadison would receive in state at minnesota looking at Chicago area schools applying early decision to UMN

Retake

Once Minnesota adcoms wake up and accept the massive hit they took last year, you'll probably be in with a 165. Keep up the pressure with the Dean. Still, its a lot of debt to assume for the prospects. A 167 would net you a very large scholarship at the T18-25 range schools, but I understand that's not always the most desirable response.

Dont go to Chicago-Kent regardless of scholarship

I wouldn't apply ED to UMN since you're above the new median.

Deleted, because law school admission probably do not fit a normal distribution. See http://admissionsbythenumbers.blogspot.com/.

Chances are you should retake.

Arcticlynx wrote:UMN - 33.8% ED is a good decision
UWisconsin - 50.2% if you can get in-state in Wisconsin after going to undergrad there.

I don't think you're likely to find much money at either of these schools though.
Based on a regression model.

How can you incorporate the new median in this model? I'm assuming you are relying on outdated numbers here, or you've installed an RNG that could be equally chaotic...

Deleted, because law school admission probably do not fit a normal distribution. See http://admissionsbythenumbers.blogspot.com/.

If you want to work in Chicago, you need to retake for a chance at NU.

Don't ed to gulc down. Not worth sticker

Arcticlynx wrote:I built a regression model in the statistical program "R" in order to help myself decide which schools to apply to. The numbers that I am using for Minnesota bases set are LSAT 156/164/167, GPA 3.44/3.79/3.90, which I believe are the most recently published for the the class of 2016. The program fits these numbers using a least squares algorithm and then extrapolates based on a two-factor GPA/LSAT score.

I don't know what an RNG is??

Random number generator. That seems to be what you are doing to fit data to the new medians. You aren't working with an actual sample, just a fictitious number generator based on quartile values.

I'm not sure how illustrative that is going to be for OP, or more generally, for you. Most schools don't operate on that sort of predictive model. UMN is very splitter friendly - if you applied actual admissions data (or regressed using the admitted students) to the medians for 2015, you would see a whole bunch of accepted students with 166-168 with GPA's far below the 25th.

The question is less statistical, and more theoretical/psychological in nature: will Minnesota accept 164 as a new median, and begin to admit 164 and 165's with low GPA's accordingly? If they repeat their previous years' practice but with this new number, OP stands far higher than 33% chance at UMN and the model is useless. If their admissions remains in denial and accepts only 166+, then the 33% will mean even less. See what I am saying? Or maybe I am misunderstanding what it is exactly that you are doing in R?

Deleted, because law school admission probably do not fit a normal distribution. See http://admissionsbythenumbers.blogspot.com/.

Arcticlynx wrote:Not using a random number generator, but I am modeling on quartiles so your critique is credited. I am using quartiles to fit a function to Minnesota's data by lest squares and then back-feeding OPs numbers into that function in order to calculate OP relative to the Class of 2016 a UMN.

Have you done any tests to see how well your method stacks up against real data?

Thank you for the feedback guys, I'm wondering why even if I am above the new medians why you would advise against applying ED? I'm assuming it can only help my chances, but I'm wondering if it will limit my ability to get a scholly. Also, what's the general consensus on whether it would it be worth it to go to a lower ranked school (William Mitchell, Loyola, DePaul, etc) with a sizable scholarship than attend UMN at/near sticker price?

Deleted, because law school admission probably do not fit a normal distribution. See http://admissionsbythenumbers.blogspot.com/.

Threezus wrote:Thank you for the feedback guys, I'm wondering why even if I am above the new medians why you would advise against applying ED? I'm assuming it can only help my chances, but I'm wondering if it will limit my ability to get a scholly. Also, what's the general consensus on whether it would it be worth it to go to a lower ranked school (William Mitchell, Loyola, DePaul, etc) with a sizable scholarship than attend UMN at/near sticker price?

Minnesota is not a very good school if you want to be employed as a lawyer. It's an especially poor school if you want to be employed as a lawyer and pay off BIG DEBT. EDing to a school like that and being locked into paying (instate) sticker is a no go amigo.

Retake

BigZuck wrote:

Threezus wrote:Thank you for the feedback guys, I'm wondering why even if I am above the new medians why you would advise against applying ED? I'm assuming it can only help my chances, but I'm wondering if it will limit my ability to get a scholly. Also, what's the general consensus on whether it would it be worth it to go to a lower ranked school (William Mitchell, Loyola, DePaul, etc) with a sizable scholarship than attend UMN at/near sticker price?

Minnesota is not a very good school if you want to be employed as a lawyer. It's an especially poor school if you want to be employed as a lawyer and pay off BIG DEBT. EDing to a school like that and being locked into paying (instate) sticker is a no go amigo.

Retake

Especially since he wants (seemingly?) to work in Chicago.

Anecdote: My used car salesman (Honda) gave me a lecture about how hard it would be to go to law school if I didn't have good grades. He was a William Mitchell grad. I did not have good grades, and I'm at a T18 now.

Arcticlynx wrote:Not using a random number generator, but I am modeling on quartiles so your critique is credited. I am using quartiles to fit a function to Minnesota's data by lest squares and then back-feeding OPs numbers into that function in order to calculate OP relative to the Class of 2016 a UMN.

If I have this correct, you're essentially estimating an index formula and then placing OP on it? That would be fine if admissions were done on a sliding scale. Unfortunately, it's not. It's done by medians, so that index is not a useful or even particularly meaningful metric.

cotiger wrote:

Arcticlynx wrote:Not using a random number generator, but I am modeling on quartiles so your critique is credited. I am using quartiles to fit a function to Minnesota's data by lest squares and then back-feeding OPs numbers into that function in order to calculate OP relative to the Class of 2016 a UMN.

If I have this correct, you're essentially estimating an index formula and then placing OP on it? That would be fine if admissions were done on a sliding scale. Unfortunately, it's not. It's done by medians, so that index is not a useful or even particularly meaningful metric.

Yes it is a correlation index, based on a matrix of two covariant factors LSAT and GPA relative to a nonlinear model (see multifactor nonlinear regression). I guess you could call it a sliding scale, but in 3 dimensions.

However, as http://admissionsbythenumbers.blogspot.com/ points out, law schools don’t act like most physical variables (it’s not like predicting wind-speed or molecular activity) and probably doesn't fit a normal distribution. So I am going to delete these posts, because without an understanding of the model and it’s limitations they are very likely to be misinterpreted and become counterproductive.