Law School Discussion

Applying to Law School => Law School Admissions => Topic started by: TheRealMongoose on February 01, 2008, 09:07:30 PM

Title: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 09:07:30 PM
Hi there,

I know some of my users/supporters have probably made a bad name for me on this site, but I was wondering if you could help me.

My law school admissions calculator is up and running (www.lawschoolmongoose.com/admissions-calculator.html) and I was hoping some of you could check it out and report back with any problems/inaccuracies.  If you've looked at it before please know that I've made a few changes in the last hour or so that should correct some of the problems we had before.

The data is calculated using the 25th/75th percentiles for both GPA and LSAT.  Medians are off because we simply took the mean of both the 25th and 75th.  This is more to save time so we don't have to search for the medians on each law school website each year.  The difference between real and calculated shouldn't make too much of a difference though.

The actual algorithm is is sort of complicated, but it weighs LSAT and GPA differently, calculates an index and then makes adjustments to it based on a number of rules.

Any suggestions would be great.  I'm not doing this for myself, only to help students next cycle, so go easy on me - I'm really not a bad guy :).  Thanks in advance and good luck with your apps!
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 09:24:34 PM
OK thanks!  That's what I want... what are your stats (if you don't mind)?

Also, the top 3 are really hard to get right without making exceptions for them.  I'm happy erring on the side of giving someone hope, though... but maybe I can fine-tune it a little.
Title: Re: Any advice for my graph?
Post by: boilerguy84 on February 01, 2008, 09:34:20 PM
i think one improvement would be if you could get the schools all the way down through #100. i know many of the people on here are only interested in t14..but there ARE others who are looking at schools lower in the top 100.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 09:36:12 PM
Thanks - there is a "next" button at the top, but it's not very clear (you're not the only one who didn't see it).  I'm going to see if I can get more on one page, though.
Title: Re: Any advice for my graph?
Post by: hbb on February 01, 2008, 09:37:36 PM
Although I think you underestimate the impact using the quartiles to generate a mean may have on your data, I do think it's pretty cool to see a positive association between GPA and LSAT at the top ranked schools, and a negative association at the lowest ranked schools.

Consider altering the scales somewhat - no need to have a scale that is so much larger than the data range.

Also, why 4 separate charts based upon rankings? Is there some sort of limit to the data points you can enter in the software you are using?

Quote
The actual algorithm is is sort of complicated, but it weighs LSAT and GPA differently, calculates an index and then makes adjustments to it based on a number of rules.

What in the world do you mean? How and why are you weighting the results? Something is very wrong: a 4.0 with a 170 is considered a reach (yellow) at schools 62-67 (which shows 10 data points? for 5 schools? WTF?) even though this is clearly a lock at all of these schools based upon the averages you've provided.

Give me a bit an I'll see if I can figure out a way for you to use the quartile data without screwing it up...

 

edit - Ah - several are tied in the rankings, leading to more schools than 5. Also, it may just be that the color is off - when moused over, a "Mongoose says you're in!" pops up.



Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 09:38:02 PM
Quote
i would say lose the graph, it's kind of annoying to use. instead, output in text form.

Thanks - maybe I'll look into doing both.  I think the graph is nice to quickly see where your chances fall based on the rank of schools, but I agree text output would be nice.  I've just seen a lot of text outputs and figured this would be a little different.
Title: Re: Any advice for my graph?
Post by: boilerguy84 on February 01, 2008, 09:41:21 PM
Thanks - there is a "next" button at the top, but it's not very clear (you're not the only one who didn't see it).  I'm going to see if I can get more on one page, though.

oh, i saw it. i just want to see some of the schools even lower (down to 100 or so). i like the format of it. i think it's really easy to understand what it's trying to tell the user. i'm not sure about the accuracy (i'm a splitter...so nothing is really all that accurate for me, anyway). if it's accurate, and has more schools, i think it can be a fantastic calculator because of the ease of use.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 09:43:56 PM
Quote
Although I think you underestimate the impact using the quartiles to generate a mean may have on your data, I do think it's pretty cool to see a positive association between GPA and LSAT at the top ranked schools, and a negative association at the lowest ranked schools.

Consider altering the scales somewhat - no need to have a scale that is so much larger than the data range.

Also, why 4 separate charts based upon rankings? Is there some sort of limit to the data points you can enter in the software you are using?

Quote
The actual algorithm is is sort of complicated, but it weighs LSAT and GPA differently, calculates an index and then makes adjustments to it based on a number of rules.

What in the world do you mean? How and why are you weighting the results? Something is very wrong: a 4.0 with a 170 is considered a reach (yellow) at schools 62-67 (which shows 10 data points? for 5 schools? WTF?) even though this is clearly a lock at all of these schools based upon the averages you've provided.

Give me a bit an I'll see if I can figure out a way for you to use the quartile data without screwing it up...

Thanks a ton.

The data is 20 per page because we were playing with "Median Scores by Rank" section and 20 was all that it could handle.  I agree that we need more per page.

The weighting just means that a median LSAT isn't equal to a median GPA.  Someone with a median LSAT and a low GPA has a better chance than someone who has a median GPA and low LSAT at most schools. I'm not getting the 4.0/170 results you are.  Let me know if you keep getting them.

Thanks so much for your input.
Title: Re: Any advice for my graph?
Post by: PABitz on February 01, 2008, 09:44:45 PM
I know this would be a bit of extra work, but you can get the actual medians from the LSAC site (ABA data sheets).
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 09:45:12 PM
Quote
oh, i saw it. i just want to see some of the schools even lower (down to 100 or so). i like the format of it. i think it's really easy to understand what it's trying to tell the user. i'm not sure about the accuracy (i'm a splitter...so nothing is really all that accurate for me, anyway). if it's accurate, and has more schools, i think it can be a fantastic calculator because of the ease of use.

Thanks, I forgot to mention that I haven't added all the data.  It will be there in the next few days or so.
Title: Re: Any advice for my graph?
Post by: hbb on February 01, 2008, 09:49:44 PM
Quote
I'm not getting the 4.0/170 results you are.

Take another look at my post, I edited and added more info about this.


So, the weighting is not affecting the placement of an individuals data in the scatter plot, but is affecting the reach/target value?


Title: Re: Any advice for my graph?
Post by: PABitz on February 01, 2008, 09:51:05 PM
Also, just a cosmetic thing. I think it would be better without the Mongoose says... notes when hovering over a school (just have the school name and medians). I know this was originally there because you didn't have the key, but now that you can tell the difference between a target, reach, etc. without hovering, the comments just add clutter.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 10:11:04 PM
Quote
Take another look at my post, I edited and added more info about this.


So, the weighting is not affecting the placement of an individuals data in the scatter plot, but is affecting the reach/target value?

Correct, the weighting only affects the calculation, not the where the data is.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 10:11:32 PM
Quote
Also, just a cosmetic thing. I think it would be better without the Mongoose says... notes when hovering over a school (just have the school name and medians). I know this was originally there because you didn't have the key, but now that you can tell the difference between a target, reach, etc. without hovering, the comments just add clutter.

Good call, I'll change that.
Title: Re: Any advice for my graph?
Post by: MorningLattes on February 01, 2008, 10:24:34 PM
boo! I want a calculator for URMs. lol...  ;)

but seriously, sucks not to have a realistic calculator for us URMs- hence having to apply to almost 20 schools cuz you just never know.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 10:32:09 PM
^^^ I'm working on it.

What sort of numerical boost do you think a URM gets?  20%?

My guess is that URM boost increases with rank and drops near the top (where there are so many qualified applicants that a URM gets a small boost, but not a huge one).  Also, it will be difficult to incorporate for schools in the midwest where URM's get a huge boost (because there are so few URM applicants).

Any thoughts?

Title: Re: Any advice for my graph?
Post by: Pop Up Video on February 01, 2008, 10:38:13 PM
HYS are not targets for me. It seems pretty off IMO.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 11:49:16 PM
Quote
HYS are not targets for me. It seems pretty off IMO.

I would disagree - but even so, like I said, the top three schools are difficult to get perfect.  In any other circumstance with a different school, being at or above the 75th LSAT percentile would make the school a target.  Like I said though, it's difficult to get HYS working properly (compared to other schools) because of their selectivity and the huge number of qualified applicants.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 01, 2008, 11:51:25 PM
Quote
i would think this so difficult to gauge that it may not be worth doing, maybe just put a little disclaimer at the bottom. i guess you could spend some time on LSN and figure out if there are any norms. however, i think it's hard to judge because i think they care about your story too. if being a URM involves overcoming hardship, etc. and you write a good PS/DS then i think you get much more of a boost. if you're just a black kid from a middle class home with average application materials then i don't think it will be as helpful. while the process is very numbers oriented adcomms do still try to put together a class full of diverse viewpoints. that's why the URM (with overcoming adversity) helps.

I agree - it sucks because being a URM is almost always a rather large boost, but it's so difficult to measure.
Title: Re: Any advice for my graph?
Post by: limegreen on February 02, 2008, 09:51:54 AM
It says for me that USC, UCLA, UT, Vandy, etc. are targets and I got WL at Vandy, I figured those were all big reaches.
Title: Re: Any advice for my graph?
Post by: _retired_ on February 02, 2008, 11:26:27 AM
Worked fine for me, with the exception of Stanford and Boalt.  They came out at target and likely respectively, when it should really be reach and target.  Can't really expect otherwise though, considering how atypical their admissions practices are.

ETA:  Although you might want to consider weighting GPA more heavily at Stanford, might help a bit.
Title: Re: Any advice for my graph?
Post by: JeNeSaisLaw on February 02, 2008, 11:35:43 AM
This thing gets me wrong entirely.

It tells me that Northwestern is a great target school, along with GULC, Berkeley, Duke, and  Michigan. There's almost no use of "reach" for me. It says all of those are targets, then that UVa (which should be the same as GULC I think) is a reach, to long shots and prayers.

I think you need a few definitions, at least for the tiptop schools, for noticeable trends. For instance, at GULC anything below 170 better have a great GPA. Duke doesn't seem to really like splitters unless you have a really high LSAT. Berkeley loves their high GPAs. NU also loves their 170+s with a low GPA, but only where there's work experience. Without taking the particular schools (at least at the top) patterns into consideration, you cannot make this board happy.
Title: Re: Any advice for my graph?
Post by: OnTheRoad on February 02, 2008, 12:39:33 PM
I'm into this. I'm really not a fan of the one over on Chiasu, not sure why, it's just too complicated to follow, a little too hopeful, etc. The thing about HYS is, how can anyone know what they're really looking for. Especially HS, you've got no idea what they'll do with you whether you're 4.0/178 or 3.2/165. So you can't expect a calculator to figure it out. I give the graph good marks except that it said I'd be good to go at UCLA and Vandy and instead they YPed me. Or rejected me. But obviously I prefer to think one over the other.
Title: Re: Any advice for my graph?
Post by: _retired_ on February 02, 2008, 01:27:44 PM
I'm into this. I'm really not a fan of the one over on Chiasu, not sure why, it's just too complicated to follow, a little too hopeful, etc. The thing about HYS is, how can anyone know what they're really looking for. Especially HS, you've got no idea what they'll do with you whether you're 4.0/178 or 3.2/165. So you can't expect a calculator to figure it out. I give the graph good marks except that it said I'd be good to go at UCLA and Vandy and instead they YPed me. Or rejected me. But obviously I prefer to think one over the other.

You mean YS?
Title: Re: Any advice for my graph?
Post by: OnTheRoad on February 02, 2008, 05:48:05 PM
No I mean HS. I think Yale is a little more predictable about their class. I wouldn't apply there without insane stats.
Title: Re: Any advice for my graph?
Post by: _retired_ on February 02, 2008, 06:11:20 PM
No I mean HS. I think Yale is a little more predictable about their class. I wouldn't apply there without insane stats.

Not that I'd know (having neither applied at nor researched HYS), but I always thought that Harvard became pretty easy to predict around 3.8+/174+ while Yale and Stanford could virtually never be guaranteed.
Title: Re: Any advice for my graph?
Post by: OnTheRoad on February 02, 2008, 06:13:21 PM
No I mean HS. I think Yale is a little more predictable about their class. I wouldn't apply there without insane stats.

Not that I'd know (having neither applied at nor researched HYS), but I always thought that Harvard became pretty easy to predict around 3.8+/174+ while Yale and Stanford could virtually never be guaranteed.

Well that's a solid point. I was thinking in terms of the other end, how far down they'd go.
Title: Re: Any advice for my graph?
Post by: _retired_ on February 02, 2008, 06:16:09 PM
No I mean HS. I think Yale is a little more predictable about their class. I wouldn't apply there without insane stats.

Not that I'd know (having neither applied at nor researched HYS), but I always thought that Harvard became pretty easy to predict around 3.8+/174+ while Yale and Stanford could virtually never be guaranteed.

Well that's a solid point. I was thinking in terms of the other end, how far down they'd go.

Which in retrospect is probably more useful, because a 3.8/174 doesn't need a graph to figure out where to apply.  :)
Title: Re: Any advice for my graph?
Post by: OnTheRoad on February 02, 2008, 06:19:05 PM
No I mean HS. I think Yale is a little more predictable about their class. I wouldn't apply there without insane stats.

Not that I'd know (having neither applied at nor researched HYS), but I always thought that Harvard became pretty easy to predict around 3.8+/174+ while Yale and Stanford could virtually never be guaranteed.

Well that's a solid point. I was thinking in terms of the other end, how far down they'd go.

Which in retrospect is probably more useful, because a 3.8/174 doesn't need a graph to figure out where to apply.  :)

I think the only thing a 3.8/174 needs is a list of demands to send to schools.
Title: Re: Any advice for my graph?
Post by: laughing_heir on February 03, 2008, 03:28:45 AM
Does "splitter" mean my stats come from different planets?  That's me. I'm 167 / 2.9. As for soft stats, I'm white, male, hardly volunteered and not even gay (worst luck), so your calculator should have all it needs and I think it hit the targets just right. One was WaUSL, which I hadn't originally considered because of my stinky GPA--but they enticed me with a fee waiver, and sure enough I was accepted. 

It kind of blew up on the reaches, though.  Southern Cal? Boalt?  Surely ye jest.  I did not even allow them to spit in my general direction.

But I like the concept a lot.  I say keep the colors and the graph format.

Back to reaches: LSAC's predictor named Illinois, Ohio, Colorado, Fordham, Maryland (all just under 50/50 odds).  If I'm a USNWR true believer, why would I reach for these reaches?  I don't think I have reaches that outrank Washington in USN.  Everything that outranks Washington is out of my reach, I think.  My reaches are (slightly) lower ranked schools. 

As it happens I'm not a true believer and I did apply to Illinois.
Title: Re: Any advice for my graph?
Post by: JeNeSaisLaw on February 03, 2008, 10:28:28 AM
Interesting breakdown in functionality when you plug in 165/3.03. It says 1-67 are a prayer. That's definitely wrong and not even close.
Title: Re: Any advice for my graph?
Post by: TheRealMongoose on February 03, 2008, 10:47:43 AM
^^^^ You probably entered the numbers in the wrong spots.  It works fine for me.
Title: Re: Any advice for my graph?
Post by: JeNeSaisLaw on February 03, 2008, 10:57:44 AM
*&^%, good call. My bad.
Title: Re: Any advice for my graph?
Post by: OlderandWiser on February 03, 2008, 11:17:05 AM
Pretty good predictor for me thus far.  The only minor difference was that it gave UIUC as a target (dinged almost immediately, I think I must have called the Dean an ahole in my PS by accident) and UWashington as a reach (accepted).  But everything else matched up right.  I applied all over the place hoping my solid soft factors might get me admitted somwhere surprising.  I'll post again when I'm further along in the cycle.  But overall I like the visual tool.  Very easy to use and a good way to get an idea at a quick glance.