Top Law Schools

Historically, the law school predictor has been said to be pretty terrible for splitters and I assume worse for super splitters, due to their use of index formulas to predict chances of admission.

They updated to a new version in august, anyone know anything about this or feel like it's suddenly better for splitters? They said the update uses a new admissions index formula, but i have no idea what this means.

Thoughts? I kind of remember reading that the LSP guy is on TLS, maybe he will see this and answer. Anyone else?

My chances definitely changed since the last update, but they seem more optimistic, which is the opposite result i'd expect if the update was more accurate for splitters as my experience has been that LS is too generous for splitters.

Is there a better predictor for super splitters? I can't find one, and lsn is pretty much zero help. I know none are going to be particularly accurate, but is one slightly more accurate than the others?

thanks!

use mylsn (actual info)

the predictor is worthless and always has been for all applicants

^

wolfgang wrote:and lsn is pretty much zero help

Yeah, predictors are tough. The problem is that they depend on data, but there's not much out there for big splitters. Who knows what's going to happen this cycle? It's gonna be insane, hopefully.

http://mylsn.info/

If you're such a splitter that there's few data points, just do the automatic search. It'll expand the range around your scores until there's enough data to be relevant. This is going to be the best you can do.

Thanks for the replies.

I think I was pretty clear that I am already well aware of lsn (including mylsn, which is just the same data).

I'm such a splitter that I'm a sub 3.0, and so there are about 5 data points at most for schools within the top 14. Plus, I'm right at a 2.9 so expanding the search doesn't really help, as it bumps the search above a 3.0 and GROSSLY overestimates my chances.

But, none of this was really the point, although I do appreciate the responses.
Perhaps I should have distilled the post down a little: I was wondering two things, for the most part.

1. Does anyone know if LSP has changed enough that is becomes useful to splitters? It's okay if nobody knows.

2. Is there a better PREDICTOR (i.e. something that uses some statistical model, as opposed to just data points) for splitters? I'm not even necessarily interested in a great one, just wondering if one is better than the rest, even marginally. Again, if nobody knows, that's fine!

Thanks again

wolfgang wrote:Thanks for the replies.

I think I was pretty clear that I am already well aware of lsn (including mylsn, which is just the same data).

I'm such a splitter that I'm a sub 3.0, and so there are about 5 data points at most for schools within the top 14. Plus, I'm right at a 2.9 so expanding the search doesn't really help, as it bumps the search above a 3.0 and GROSSLY overestimates my chances.

But, none of this was really the point, although I do appreciate the responses.
Perhaps I should have distilled the post down a little: I was wondering two things, for the most part.

1. Does anyone know if LSP has changed enough that is becomes useful to splitters? It's okay if nobody knows.

2. Is there a better PREDICTOR (i.e. something that uses some statistical model, as opposed to just data points) for splitters? I'm not even necessarily interested in a great one, just wondering if one is better than the rest, even marginally. Again, if nobody knows, that's fine!

Thanks again

Humor me for a second. Look at Northwestern on this chart: http://mylsn.info/we612h
The range is 2.77-3.03 for GPA, which shouldn't skew high for your 2.9. Georgetown is over the same range. Plus, it's over the last three cycles, and this cycle should be more favorable. Granted, it's only 8 data points, but still, for those 8 there are five acceptances, two waitlists, and only one rejection.

As for other models, no, I'm not aware of any.

yeah. The only problem with NU is that their numbers will obviously be pretty similar in that range, because a) they are splitter-friendly and care about work experience, and seemed to have no GPA floor for 170+ applicants last cycle.

So their 2.99/3.0 distinction will definitely be less pronounced than other schools. But in a lot of cases, getting over 3.0 changes things, and the auto search doesn't necessarily reflect that.

Maybe my numbers are just weird. For example, my gpa is 2.9x. If I search from 2.0-2.99, and put a reasonable range for my lsat (i.e. I chose the range in between 170-180 that give me the best chances when coupled with a 2.0-2.99 just to be safe) I get certain chances.

When I put in my lsat and gpa in the auto search, a few of them go up just a couple of percentage points. Some of them stay the same. I assume these are insignificant.

THe rest jump A LOT. for example, it goes from 7% to 20% at penn, 65 to 78% at northwestern, from 0 to 13% at cornell, from 29%-63% at gulc, from 24-60% at vanderbilt, etc.

I always hear there's a huge psychological bias when you break 3.0, and that certainly seems to be the case here.

Of course, I'd love to hear that the auto search when you have a 170+ and a 2.9x is legitimate and that the 2.99/3.0 distinction is disappearing. Heck, a 63% shot at gulc? I'll take that action any day

I hate that there's no predictor for splitters! As if things aren't stressful enough haha.

Thanks for all the replies!

wolfgang wrote:yeah. The only problem with NU is that their numbers will obviously be pretty similar in that range, because a) they are splitter-friendly and care about work experience, and seemed to have no GPA floor for 170+ applicants last cycle.

So their 2.99/3.0 distinction will definitely be less pronounced than other schools. But in a lot of cases, getting over 3.0 changes things, and the auto search doesn't necessarily reflect that.

Maybe my numbers are just weird. For example, my gpa is 2.9x. If I search from 2.0-2.99, and put a reasonable range for my lsat (i.e. I chose the range in between 170-180 that give me the best chances when coupled with a 2.0-2.99 just to be safe) I get certain chances.

When I put in my lsat and gpa in the auto search, a few of them go up just a couple of percentage points. Some of them stay the same. I assume these are insignificant.

THe rest jump A LOT. for example, it goes from 7% to 20% at penn, 65 to 78% at northwestern, from 0 to 13% at cornell, from 29%-63% at gulc, from 24-60% at vanderbilt, etc.

I always hear there's a huge psychological bias when you break 3.0, and that certainly seems to be the case here.

Of course, I'd love to hear that the auto search when you have a 170+ and a 2.9x is legitimate and that the 2.99/3.0 distinction is disappearing. Heck, a 63% shot at gulc? I'll take that action any day

I hate that there's no predictor for splitters! As if things aren't stressful enough haha.

Thanks for all the replies!

If you're concerned there's a psychological thing at 3.0, then just search for 2.75-2.99. That, coupled with 170-176 for the last three cycles shows that Northwestern and GULC are absolutely possible for you in the T14. Vandy's also a good shot, and apply to WUSTL for guaranteed admission with $$.

Tweak the numbers however you want, but as the sample size is small it's going to be noisy, so don't worry about the actual percentages. <25% acceptance means you've got to be reeeally special, 25%-75% means you've got a shot, and 75%+ means you're most likely in. So like I said, you're a shoo-in at WUSTL; Vandy, NU, and Georgetown are get-able; and there are a couple others that might be worth sending an app to if you have a really good reason for wanting to go there. Unfortunately, that's as precise as you're going to get.

wolfgang wrote:Thanks for the replies.

I think I was pretty clear that I am already well aware of lsn (including mylsn, which is just the same data).

I'm such a splitter that I'm a sub 3.0, and so there are about 5 data points at most for schools within the top 14. Plus, I'm right at a 2.9 so expanding the search doesn't really help, as it bumps the search above a 3.0 and GROSSLY overestimates my chances.

But, none of this was really the point, although I do appreciate the responses.
Perhaps I should have distilled the post down a little: I was wondering two things, for the most part.

1. Does anyone know if LSP has changed enough that is becomes useful to splitters? It's okay if nobody knows.

2. Is there a better PREDICTOR (i.e. something that uses some statistical model, as opposed to just data points) for splitters? I'm not even necessarily interested in a great one, just wondering if one is better than the rest, even marginally. Again, if nobody knows, that's fine!

Thanks again

I just ran my numbers through (I was a splitter as well, but not as extreme as you so there are plenty of LSN data points to compare to) and it's still utterly worthless.

I know of no other "predictors" except for the LSAC one which is if anything even worse (caveat: I don't know if it's based on statistics or data, but that distinction is irrelevant considering it's also useless).

ETA: At least it doesn't tell me I would have had a better shot at SLS than CLS anymore