Law School Predictor: The Thread" Forum

[YCrevolution]

..

[superflush]

I gotta wait until i get home to view this on a bigger monitor. Should be resourceful though

[YCrevolution]

..

[YCrevolution]

..

[superflush]

I gotta wait until i get home to view this on a bigger monitor. Should be resourceful though

Oh, I'm home now, so I'll pull this up on my big monitor.

[Helmholtz]

While I have been working on creating predictions for Canadian law schools, I think I'm going to put that feature on hold until I become more comfortable with the prediction being rendered. Right now, I don't have much data to go off of to make sure the formulas aren't too out-of-sync and it's hard to find 25/75 data for Canadian law schools (they have median data, but I'll have to do some more adaptation to make that work).

On the other hand, I've also been working on decreasing some overly optimistic predictions for applicants with an okay (or better) LSAT but a not-so-good GPA. My plan is to implement a prediction penalty for applicants who fall below either 3.0 or a set amount below a school's 25%, whichever is lower, and which increases depending on far below 3.0 or 25% the applicant's GPA falls. My goal is to introduce a feature that better reflects the notion that sometimes even a good LSAT will have trouble erasing a poor GPA.

Canadian law schools don't really exist. It's just propaganda by the Canadian government to make them appear less lame.

[YCrevolution]

..

[Iuvo]

Congrats!

[YCrevolution]

..

[SilverE2]

I recently searched for "law school predictor" on Google (without the quotes), and it's become the top Google search result (beating out the official LSAC predictor and Chiasu). All in one month's time. It wouldn't have been possible with the support and publicity provided by TLS forum members.

Congrats . Now your next step is to have it so when you search for "law school", your site is the top google search, and then just "law". Haha, you have quite a ways to go .

[pissantvache]

Quick question: although your predictor clearly doesn't account for it, do you imagine that it would be significant to factor in the frequency of both a stratospheric LSAT and a stratospheric GPA? That is, those above both the 75th percentile in LSAT and GMAT probably comprise less than 25% of the class, such that, especially at a place like Yale or HLS, comparing one's index to the indexes at 25/25 and 75/75 might distort the results, especially at the lower margins--if someone has a 30th lsat and 30th GPA, that person almost certainly isn't getting in.

Granted, the 'weak consider' result probably accounts for that, but I wonder whether even that might be optimistic in the absence of some major redeeming factors.

[YCrevolution]

..

[nickc8]

Why are some of the 25th/75th percentile LSAT numbers lower than those reported on the actual law school's website? For example, UC Davis is listed as 158/163 on your page but UC Davis claims 161/166 on their website. Do some schools attempt to "fluff" their actual numbers?

[YCrevolution]

..

[nickc8]

Good to know. I definitely appreciate the hard work you put into your website; it has helped tremendously. Thanks!

[pissantvache]

It does have some effect, most noticeably at HYS. There just isn't a particularly good formula for HYS since they reject a lot of applicants with the requisite numbers.

I would agree that splitters have a better chance when they're weak considers than others (traditional splitters do get a prediction bonus right now). Right now, I'm uncomfortable with weakening predictions for those whose numbers fall between 25% and 50% without some concrete data and evidence to suggest what kind of weakening would be appropriate..

Interesting. I agree that that's probably a good idea, and I'm definitely glad that you have the weak consider function in there. While I understand that there are probably already enough categories (dividing things in three always seems to resonate), would it perhaps be beneficial to add in a 4th, say, "Prospective Deny," that keeps people realistic without resorting to auto-deny status? or am I just being nitpicky?

[roadkilllaw]

It does have some effect, most noticeably at HYS. There just isn't a particularly good formula for HYS since they reject a lot of applicants with the requisite numbers.

I would agree that splitters have a better chance when they're weak considers than others (traditional splitters do get a prediction bonus right now). Right now, I'm uncomfortable with weakening predictions for those whose numbers fall between 25% and 50% without some concrete data and evidence to suggest what kind of weakening would be appropriate..

Interesting. I agree that that's probably a good idea, and I'm definitely glad that you have the weak consider function in there. While I understand that there are probably already enough categories (dividing things in three always seems to resonate), would it perhaps be beneficial to add in a 4th, say, "Prospective Deny," that keeps people realistic without resorting to auto-deny status? or am I just being nitpicky?

You're just being nitpicky!

[YCrevolution]

..

[BaiAilian2013]

I just plugged my roommate's stats in for fun, and it came up with Deny for literally every school that she was accepted to during this past cycle. Is that a mark of how much harsher admissions will be this cycle ( ), or are there factors that are known to make the predictor less accurate (e.g. URM, strong upward trend in GPA)?

Edit for clarity: I do see the predictor has a URM box and I did check it; I'm just flailing around for explanations for the discrepancy.

[YCrevolution]

..

[jackassjim]

Is that a mark of how much harsher admissions will be this cycle ( )

It can't be since much of the formula is based on data from past years (this is especially true for schools that don't publish an official index formula).

[tarheel87]

For the "From Middle" index numbers, is number like -0.08 better than -10.1, basically what is a good number for that. I don't really "get" indices.

[YCrevolution]

..

[Lmao Zedong]

.

[Capercaillie]

Edit: You owe me, Garg.