Law School Predictor: The Thread" Forum

[hypermeganet]

YC, have you looked into incorporating LSAT cutoffs into your predictor?

The predictor still seems to favor reverse splitters quite a bit when, as we all know, that they are really at a huge disadvantage in the process when compared to "normal" splitters.

A 162/4.0 gives you "Strong Consider" at USC and ND and admits at BC and Emory. Based on LSN for the last cycle, not one person was admitted to any of those schools with those numbers (URMs excluded); just rejects and WLs. The LSAT is just too far off of their median, even if the overall index works out to be fairly high.

[keg411]

I was going to ask about that as well. I put my friends numbers in for fun and it gave her extremely optimistic chances at a ton of schools. Whatever formula is being used now LOVES high GPA splitters and appears to give a much stronger weight to GPA than LSAT.

[YCrevolution]

..

[summerschooled]

I was going to ask about that as well. I put my friends numbers in for fun and it gave her extremely optimistic chances at a ton of schools. Whatever formula is being used now LOVES high GPA splitters and appears to give a much stronger weight to GPA than LSAT.

It doesn't like high GPA splitters anymore than it used to (at least from Version 2.5 to 2.6).

There are plans to take into account school-specific LSAT and GPA cutoffs/walls in a near-future version (2.7 or 3.0). A state residency feature is also under development.

Any idea on how skewed it might be for high-LSAT, low-GPA people?

[citydweller09]

.

[YCrevolution]

..

[YCrevolution]

..

[Bronte]

I asked this once before and I guess the question got lost in the shuffle, so I'll give it another go:

Are there any plans to differentiate between different types of URMs in future versions? The difference in the boost received by, say, a Mexican female is substantially different than that received by an AA male (the holy grail of URM-hood ...). I ask because as it stands now, the predictor seems pretty pessimistic about URM admissions, and this might be a factor.

While I'm open to doing so, I'd need to collect enough data to make such differentiations. Right now, I don't have that data.

There's a lot of things LSN could do for next cycle that would make it a way better product in terms of the data collected. Who runs that site anyway?

[summerschooled]

I was going to ask about that as well. I put my friends numbers in for fun and it gave her extremely optimistic chances at a ton of schools. Whatever formula is being used now LOVES high GPA splitters and appears to give a much stronger weight to GPA than LSAT.

It doesn't like high GPA splitters anymore than it used to (at least from Version 2.5 to 2.6).

There are plans to take into account school-specific LSAT and GPA cutoffs/walls in a near-future version (2.7 or 3.0). A state residency feature is also under development.

Any idea on how skewed it might be for high-LSAT, low-GPA people?

Not much, at least on the aggregate level. It's usually just a few percentage points lower than the prediction categories would suggest (except for extreme splitters ala 170/2.0-style, whose "Deny" predictions are sometimes overly pessimistic).

Okay, thanks!

[LSATfromNC]

I asked this once before and I guess the question got lost in the shuffle, so I'll give it another go:

Are there any plans to differentiate between different types of URMs in future versions? The difference in the boost received by, say, a Mexican female is substantially different than that received by an AA male (the holy grail of URM-hood ...). I ask because as it stands now, the predictor seems pretty pessimistic about URM admissions, and this might be a factor.

While I'm open to doing so, I'd need to collect enough data to make such differentiations. Right now, I don't have that data.

There's a lot of things LSN could do for next cycle that would make it a way better product in terms of the data collected. Who runs that site anyway?

Not sure, but I wish they would update the rankings and school profiles.

[Bronte]

I asked this once before and I guess the question got lost in the shuffle, so I'll give it another go:

Are there any plans to differentiate between different types of URMs in future versions? The difference in the boost received by, say, a Mexican female is substantially different than that received by an AA male (the holy grail of URM-hood ...). I ask because as it stands now, the predictor seems pretty pessimistic about URM admissions, and this might be a factor.

While I'm open to doing so, I'd need to collect enough data to make such differentiations. Right now, I don't have that data.

There's a lot of things LSN could do for next cycle that would make it a way better product in terms of the data collected. Who runs that site anyway?

Not sure, but I wish they would update the rankings and school profiles.

Yeah, but what I'm really looking for:
(1) Allow users to show only acceptances, waitlists, or rejections on the charts (i.e., turn on/off reds, greens, and yellows)
(2) Offer links to raw data in .xls or .csv format
(3) Add better metrics that are reported in these raw data files
(4) Allow users to report certain metrics anonymously, but still report them in the raw data files
(5) Add a built in simple probability calculator (admitted over applied) that allows user to enter LSAT score and GPA range

[YCrevolution]

..

[laidoffjournalist]

Hopefully by May 25 I won't be needing it anymore... but what is "significant work experience?

[YCrevolution]

..

[los blancos]

Thanks YC... that thing has been very valuable.

I'm using a combination of LSP, hourumd, and LSN profiles to get an impression of where my 3.6/171 self will end up... hourumd is still more optimistic.

[roadkilllaw]

Thanks YC... that thing has been very valuable.

I'm using a combination of LSP, hourumd, and LSN profiles to get an impression of where my 3.6/171 self will end up... hourumd is still more optimistic.

I'd rather have a predictor be more pessimistic than more optimistic

[Carlisle]

Thanks YC... that thing has been very valuable.

I'm using a combination of LSP, hourumd, and LSN profiles to get an impression of where my 3.6/171 self will end up... hourumd is still more optimistic.

my problem with hourumd is that the it track five years. data from that long ago doesn't paint a fair picture of what your cycle will be like.

[YCrevolution]

..

[laidoffjournalist]

FYI, DePaul's link is broken (html has to Ls at the end of the URL).

[YCrevolution]

..

[laidoffjournalist]

Also, a cool idea if it's possible (and I'm not sure if it's been suggested) is if the calculator could predict the likelihood of merit pay and a ballpark of how much. It doesn't matter for me since I will most likely be finished with the application process by the time it would be implemented, but I really, really wish there was a way for me to anticipate this somewhat. To me, less debt > higher ranked school.

[YCrevolution]

..

[YCrevolution]

..

[CyLaw]

Thankfully, I passed the 1,000 post mark so I am immune to LSP.

[Haven]

Looks like youre auctioning off something!