"Section Stacking" and Class Ranking Question

A forum for applicants and admitted students to ask law students and graduates about law school and the practice of law.
User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

"Section Stacking" and Class Ranking Question

Postby ScottRiqui » Mon Dec 09, 2013 11:06 pm

I think I may have a fundamental misunderstanding about class rankings in law school. When I've read about posters here being "top 10%" or "top 20%", I've assumed that was their ranking within their entire class. But, I've also read about "section stacking", and it seems like stacking would only work if the rankings were only among the other students in your same section. And if your ranking is only within your section, wouldn't you have instances where two students in different sections could have similar GPAs, but significantly different rankings?

User avatar
Nova
Posts: 9116
Joined: Sun Apr 15, 2012 8:55 pm

Re: "Section Stacking" and Class Ranking Question

Postby Nova » Mon Dec 09, 2013 11:10 pm

youre always ranked against the entire class

but each section is curved against themselves

section staking works cause people in the stacked sections would have been statistically likely to do better in the unstacked sections.

User avatar
rinkrat19
Posts: 13909
Joined: Sat Sep 25, 2010 5:35 am

Re: "Section Stacking" and Class Ranking Question

Postby rinkrat19 » Mon Dec 09, 2013 11:13 pm

If the sections are stacked, then they put all the scholarship kids in classes together. The curve guarantees that half of them end up below median in each class, which results in a lot more of them losing their scholarships than if they were distributed evenly between sections.

09042014
Posts: 18282
Joined: Wed Oct 14, 2009 10:47 pm

Re: "Section Stacking" and Class Ranking Question

Postby 09042014 » Mon Dec 09, 2013 11:16 pm

Nova wrote:section staking works cause people in the stacked sections would have been statistically likely to do better in the unstacked sections.


But it doesn't work all that well since

1) the GPA/LSAT correlations aren't that strong

2) the average school has a pretty narrow gap. The difference between waitlist and full ride is basically within the margin of error of the GPA/LSAT correlation to LGPA.

User avatar
rinkrat19
Posts: 13909
Joined: Sat Sep 25, 2010 5:35 am

Re: "Section Stacking" and Class Ranking Question

Postby rinkrat19 » Mon Dec 09, 2013 11:23 pm

Desert Fox wrote:
Nova wrote:section staking works cause people in the stacked sections would have been statistically likely to do better in the unstacked sections.


But it doesn't work all that well since

1) the GPA/LSAT correlations aren't that strong

2) the average school has a pretty narrow gap. The difference between waitlist and full ride is basically within the margin of error of the GPA/LSAT correlation to LGPA.

I wonder if the GPA/LSAT correlations are stronger and/or the gap is larger at a shitty school (which is more likely to section stack). Cooley or whoever's class seems like it's made up of people with mediocre-to-ok numbers with huge, stip-ridden schollys, and utter mouthbreathers at full sticker.

User avatar
PepperJack
Posts: 646
Joined: Sun Sep 29, 2013 1:23 pm

Re: "Section Stacking" and Class Ranking Question

Postby PepperJack » Mon Dec 09, 2013 11:23 pm

They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high. I can also quote data on URM beneficiaries, and exam performance but that'd be going too off topic. However, the data unequivocally shows 10 point differentials are night and day in first year GPA's on average.

09042014
Posts: 18282
Joined: Wed Oct 14, 2009 10:47 pm

Re: "Section Stacking" and Class Ranking Question

Postby 09042014 » Mon Dec 09, 2013 11:25 pm

PepperJack wrote:They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high. I can also quote data on URM beneficiaries, and exam performance but that'd be going too off topic. However, the data unequivocally shows 10 point differentials are night and day in first year GPA's on average.



See the green and my point 2.

Hint: this is why I got dat 176

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Mon Dec 09, 2013 11:34 pm

rinkrat19 wrote:If the sections are stacked, then they put all the scholarship kids in classes together. The curve guarantees that half of them end up below median in each class, which results in a lot more of them losing their scholarships than if they were distributed evenly between sections.


Do the performances vary so widely between the sections that the students in the stacked section who end up below the overall class median wouldn't have likely ended up below the overall median anyway, even if they were spread into different sections? I thought the purpose of individually-curved sections was that the "median" students in all of the sections would have the same GPA?

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Mon Dec 09, 2013 11:42 pm

PepperJack wrote:They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high.


Can the bolded really be true, though? For most schools, the LSAT spread between the 25th and 75th percentiles is already 4-6 points, and that's only half of the class, not 90%.

User avatar
stillwater
Posts: 3811
Joined: Tue Jun 28, 2011 2:59 pm

Re: "Section Stacking" and Class Ranking Question

Postby stillwater » Mon Dec 09, 2013 11:45 pm

ScottRiqui wrote:
PepperJack wrote:They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high.


Can the bolded really be true, though? For most schools, the LSAT spread between the 25th and 75th percentiles is already 4-6 points, and that's only half of the class, not 90%.


because the LSAT must dramatically change after that right?????????????? some schools have the same median and 75 percentile

jarofsoup
Posts: 1951
Joined: Tue Jul 01, 2008 2:41 am

Re: "Section Stacking" and Class Ranking Question

Postby jarofsoup » Mon Dec 09, 2013 11:46 pm

First year grades are rather random. I am not sure if many respectable law schools even have a 10 point deferential in their candidates...

User avatar
Nova
Posts: 9116
Joined: Sun Apr 15, 2012 8:55 pm

Re: "Section Stacking" and Class Ranking Question

Postby Nova » Mon Dec 09, 2013 11:49 pm

jarofsoup wrote: I am not sure if many respectable law schools even have a 10 point deferential in their candidates...


Yale:152-180
http://www.law.yale.edu/admissions/profile.html

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Mon Dec 09, 2013 11:57 pm

stillwater wrote:
ScottRiqui wrote:
PepperJack wrote:They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high.


Can the bolded really be true, though? For most schools, the LSAT spread between the 25th and 75th percentiles is already 4-6 points, and that's only half of the class, not 90%.


because the LSAT must dramatically change after that right?????????????? some schools have the same median and 75 percentile


Dramatically or not, if a school's spread in the middle 50% is already 4-6 points (and several of the T-20 have even bigger spreads), then the 90% spread has to be at least that large. And with one or two notable exceptions, the schools with identical 50th and 75th numbers still have a significant spread between the 25th and the 75th.

User avatar
stillwater
Posts: 3811
Joined: Tue Jun 28, 2011 2:59 pm

Re: "Section Stacking" and Class Ranking Question

Postby stillwater » Mon Dec 09, 2013 11:59 pm

ScottRiqui wrote:
stillwater wrote:
ScottRiqui wrote:
PepperJack wrote:They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high.


Can the bolded really be true, though? For most schools, the LSAT spread between the 25th and 75th percentiles is already 4-6 points, and that's only half of the class, not 90%.


because the LSAT must dramatically change after that right?????????????? some schools have the same median and 75 percentile


Dramatically or not, if a school's spread in the middle 50% is already 4-6 points (and several of the T-20 have even bigger spreads), then the 90% spread has to be at least that large. And with one or two notable exceptions, the schools with identical 50th and 75th numbers still have a significant spread between the 25th and the 75th.


yea i agree with you

User avatar
thesealocust
Posts: 8441
Joined: Mon Oct 20, 2008 8:50 pm

Re: "Section Stacking" and Class Ranking Question

Postby thesealocust » Tue Dec 10, 2013 12:27 am

1. Section stacking is largely if not entirely a myth.

2. The rumor of section stacking is all about scholarship stipulations. Each section is grading on a curve, so the average GPA of any given section is likely to be near or exactly the average GPA of the school. The theory is that, for an extreme but obvious example, a school with 100 people and 5 sections might give 20 students scholarships requiring top third GPAs but then place half of them in each of two sections. As a result, in two section 50% of the class can now only keep their scholarship if they finish in the top third, which is nearly if not actually mathematically impossible, since each class will be graded on a curve.

As you can see, it's not really about LSAT/GPA or expected performance, it's about probability and the curve.

3. I firmly believe real section stacking like the above is rarely true, and instead crops up as myth and legend, especially at shitty schools that give you a lot of scholarship money with unrealistic stipulations. At schools that give out a lot of scholarships with stipulations, people get to talking, and suddenly realize an awful lot of people need to finish in the top X% to retain their money, and a lot are probably in EVERY section. End result, a few people find out about their scholarships in a section and declare it 'stacked.' Like the proverb: there is only one beautiful child in the world, and every mother has it.

TL;DR selective section stacking is probably false, but shitty schools with underhanded scholarship practices probably wind up giving out so many that every section may wind up believing it is a "stacked" section.

Unrelated annecdata: the person in my 1L class got the top GPA also had a bottom 25% LSAT score.

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Tue Dec 10, 2013 12:41 am

thesealocust wrote:1. Section stacking is largely if not entirely a myth.

2. The rumor of section stacking is all about scholarship stipulations. Each section is grading on a curve, so the average GPA of any given section is likely to be near or exactly the average GPA of the school. The theory is that, for an extreme but obvious example, a school with 100 people and 5 sections might give 20 students scholarships requiring top third GPAs but then place half of them in each of two sections. As a result, in two section 50% of the class can now only keep their scholarship if they finish in the top third, which is nearly if not actually mathematically impossible, since each class will be graded on a curve.

As you can see, it's not really about LSAT/GPA or expected performance, it's about probability and the curve.

3. I firmly believe real section stacking like the above is rarely true, and instead crops up as myth and legend, especially at shitty schools that give you a lot of scholarship money with unrealistic stipulations. At schools that give out a lot of scholarships with stipulations, people get to talking, and suddenly realize an awful lot of people need to finish in the top X% to retain their money, and a lot are probably in EVERY section. End result, a few people find out about their scholarships in a section and declare it 'stacked.' Like the proverb: there is only one beautiful child in the world, and every mother has it.

TL;DR selective section stacking is probably false, but shitty schools with underhanded scholarship practices probably wind up giving out so many that every section may wind up believing it is a "stacked" section.

Unrelated annecdata: the person in my 1L class got the top GPA also had a bottom 25% LSAT score.


Thanks - that makes sense. If the individual section curves work the way they're supposed to, a person who finishes 1L in the Xth overall percentile should have ended up there no matter which section they were in, right?

In your 100 students/five sections example, the odds of any 20 specific students all finishing in the top third at the end of the year were pretty negligible to begin with, no matter how they were distributed?

Basically, I keep coming back to the idea that for section stacking to work, there would have to be some flaw in how the individual sections are curved, since the ostensible purpose for the individual curves is so that you can fairly compare GPAs across sections when determining overall class rank.

User avatar
thesealocust
Posts: 8441
Joined: Mon Oct 20, 2008 8:50 pm

Re: "Section Stacking" and Class Ranking Question

Postby thesealocust » Tue Dec 10, 2013 12:42 am

Section stacking only breaks the intra-section comparison if you believe those in the stacked sections had a real leg up based on entering GPA/LSAT, which is questionable. That's not really the point. The point is that in a section where even 25% of the students need to get top third or better to keep their scholarship, it becomes mathematically very unlikely, and if there are only 0% students who need top 25% to keep their scholarship in another section, it will lead to obvious cries of shenanigans on the part of the administration.

It is generally taken for granted that overall GPA is how students are ranked and that rankings are comparable from section to section.

User avatar
PepperJack
Posts: 646
Joined: Sun Sep 29, 2013 1:23 pm

Re: "Section Stacking" and Class Ranking Question

Postby PepperJack » Tue Dec 10, 2013 12:56 am

ScottRiqui wrote:
stillwater wrote:
ScottRiqui wrote:
PepperJack wrote:They're fairly strong dude. You need to keep in mind over 90 percent of a school is within 5 LSAT points, and a 15 percent variation for under 5 LSAT points is high.


Can the bolded really be true, though? For most schools, the LSAT spread between the 25th and 75th percentiles is already 4-6 points, and that's only half of the class, not 90%.


because the LSAT must dramatically change after that right?????????????? some schools have the same median and 75 percentile


Dramatically or not, if a school's spread in the middle 50% is already 4-6 points (and several of the T-20 have even bigger spreads), then the 90% spread has to be at least that large. And with one or two notable exceptions, the schools with identical 50th and 75th numbers still have a significant spread between the 25th and the 75th.

On the 5 point variation for 90% of the class, it's mentioned here: Alex M. Johnson Jr., African Americans, Law Schools and the LSAT. Charlottesville, Virginia (UVA BLSA: February 15, 2010), available at http://www.youtube.com/watch?v=_7_xHsce57c

As for Sealocust's classmate, the 25th is still like a 165 vs a 170 median - that's 5 questions. At that range we're speaking about who found TLS, used a private tutor, took the test 3 times vs who took a Princeton course. There's no big IQ distinction. When we're speaking about schools that section stack you're comparing 160s vs 150s, and that is much more likely to produce a tangible IQ difference. I have a friend who went to a notorious section stacking school, and placed top 10%. He commented that there were literally people who had severe reading problems in his larger classes, and the reading workload was about 4-5 pages per class.

I'm not gonna post the cites because it'll create a huge thing but there is hard evidence 90% of those who get preferential admission treatment are in the bottom 10% of the law school curve. There is also hard evidence these same numerical profiles are just as likely as the rest of the class to place above median when they attend schools where their LSAT is at the median. If a school ranges between accepting candidates with a 145 and 160, that's night and day. 165 to 175, not as much.

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Tue Dec 10, 2013 12:57 am

thesealocust wrote:It is generally taken for granted that overall GPA is how students are ranked and that rankings are comparable from section to section.


If stips are based on ranking, and rankings are based on overall GPA and are comparable from section to section, doesn't that mean that the students who missed the stip cutoff in the "stacked" section likely would have missed the stip cutoff regardless of which section they were in?

It seems like the real danger is if you end up in a section full of shit-hot students, such that an objectively stellar performance that would have put you near the top of any other section actually lands you at median in your section (and therefore, median in the overall rankings).

User avatar
PepperJack
Posts: 646
Joined: Sun Sep 29, 2013 1:23 pm

Re: "Section Stacking" and Class Ranking Question

Postby PepperJack » Tue Dec 10, 2013 1:02 am

thesealocust wrote:Section stacking only breaks the intra-section comparison if you believe those in the stacked sections had a real leg up based on entering GPA/LSAT, which is questionable. That's not really the point. The point is that in a section where even 25% of the students need to get top third or better to keep their scholarship, it becomes mathematically very unlikely, and if there are only 0% students who need top 25% to keep their scholarship in another section, it will lead to obvious cries of shenanigans on the part of the administration.

It is generally taken for granted that overall GPA is how students are ranked and that rankings are comparable from section to section.

That's not really that true. There was a business started a few years back by some kids that allowed students to bet on their college grades. They'd put down a grand or so, and get 1500 if they pulled an A and break even for an A-. The business failed, because the students were just incentivized to work harder for the immediate return. Obviously, there's no forced curve here so it's not really applicable.

But problem is not only the forced curve, but the employment #'s. By putting a kid in a school with a TTTT reputation and making them compete against T1 students, they need to place close to #1 in their section. Even if they're top 25%, they're likely screwed. The school is really chipping many of these kids of the chance at gainful employment at the onset, even though the business strategy is clear. I think the schools are really in some breach of contract if they don't disclose this. Think about it. Student A takes TTTT because he gets a full ride if he's top 25%. He can get a half ride to TT with no stips. He thinks, hey median LSAT and GPA is 150 and 3.1 at TTTT. In reality, the kids he's up against are 160 and 3.5. Student A works hard, but is only top 25%. He gets no job, and took out 90k for COL and wasted 3 prime working years all based off of what may be a reckless failure to disclose.

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Tue Dec 10, 2013 1:02 am

PepperJack wrote:On the 5 point variation for 90% of the class, it's mentioned here: Alex M. Johnson Jr., African Americans, Law Schools and the LSAT. Charlottesville, Virginia (UVA BLSA: February 15, 2010), available at http://www.youtube.com/watch?v=_7_xHsce57c

As for Sealocust's classmate, the 25th is still like a 165 vs a 170 median - that's 5 questions.


I haven't watched the video yet, but if he's talking about 90% of the students being within 5 points of some particular score, that's a ten-point spread, which I could totally believe. I thought you were saying that 90% of the students were within 5 points of each other, which would only be a five-point spread (and impossible at most schools).

User avatar
PepperJack
Posts: 646
Joined: Sun Sep 29, 2013 1:23 pm

Re: "Section Stacking" and Class Ranking Question

Postby PepperJack » Tue Dec 10, 2013 1:05 am

ScottRiqui wrote:
thesealocust wrote:It is generally taken for granted that overall GPA is how students are ranked and that rankings are comparable from section to section.


If stips are based on ranking, and rankings are based on overall GPA and are comparable from section to section, doesn't that mean that the students who missed the stip cutoff in the "stacked" section likely would have missed the stip cutoff regardless of which section they were in?

It seems like the real danger is if you end up in a section full of shit-hot students, such that an objectively stellar performance that would have put you near the top of any other section actually lands you at median in your section (and therefore, median in the overall rankings).

No, you don't get how it works.

Let's say out of 10 kids, there's 5 B+'s, 2 A's, 2 B's and 1 C+.

the 2 A's are pacing a 4.0.

Doesn't matter if they're in Section xx or Section yy. I suppose if firms KNEW there was section stacking and a kid writes XXXX Scholarship on the resume, they'd know about the stacked section so wouldn't treat them the same as a student with an identical GPA not in that section. Contrarily, it'd still say Shit School XXX on their website.

User avatar
PepperJack
Posts: 646
Joined: Sun Sep 29, 2013 1:23 pm

Re: "Section Stacking" and Class Ranking Question

Postby PepperJack » Tue Dec 10, 2013 1:06 am

ScottRiqui wrote:
PepperJack wrote:On the 5 point variation for 90% of the class, it's mentioned here: Alex M. Johnson Jr., African Americans, Law Schools and the LSAT. Charlottesville, Virginia (UVA BLSA: February 15, 2010), available at http://www.youtube.com/watch?v=_7_xHsce57c

As for Sealocust's classmate, the 25th is still like a 165 vs a 170 median - that's 5 questions.


I haven't watched the video yet, but if he's talking about 90% of the students being within 5 points of some particular score, that's a ten-point spread, which I could totally believe. I thought you were saying that 90% of the students were within 5 points of each other, which would only be a five-point spread (and impossible at most schools).

Yes, within one another. I'm not writing a brief for the Supreme Court. Chill out. Where did I imply within some score?

User avatar
PepperJack
Posts: 646
Joined: Sun Sep 29, 2013 1:23 pm

Re: "Section Stacking" and Class Ranking Question

Postby PepperJack » Tue Dec 10, 2013 1:11 am

thesealocust wrote:Section stacking only breaks the intra-section comparison if you believe those in the stacked sections had a real leg up based on entering GPA/LSAT, which is questionable. That's not really the point. The point is that in a section where even 25% of the students need to get top third or better to keep their scholarship, it becomes mathematically very unlikely, and if there are only 0% students who need top 25% to keep their scholarship in another section, it will lead to obvious cries of shenanigans on the part of the administration.

It is generally taken for granted that overall GPA is how students are ranked and that rankings are comparable from section to section.

There are other factors too. A slow but deep thinker may do better on one type of exam than another. Same for the quick shallow thinker. Not being able to choose professors and exam styles really predicts much of someone's GPA too. As do simple life things like someone having the benefit of a reliable car vs someone having a piece of shit car. Someone having a breakup vs someone not having a breakup. Etc. etc. But yes, section stacking is a problem. However, I think everyone complains about their section being unfairly stacked because people delude all evidence as much as they realistically can to make themselves look as good as possible. Section stacking, if real is screwed up, particularly at non-top schools.

User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Re: "Section Stacking" and Class Ranking Question

Postby ScottRiqui » Tue Dec 10, 2013 1:15 am

PepperJack wrote:
ScottRiqui wrote:
thesealocust wrote:It is generally taken for granted that overall GPA is how students are ranked and that rankings are comparable from section to section.


If stips are based on ranking, and rankings are based on overall GPA and are comparable from section to section, doesn't that mean that the students who missed the stip cutoff in the "stacked" section likely would have missed the stip cutoff regardless of which section they were in?

It seems like the real danger is if you end up in a section full of shit-hot students, such that an objectively stellar performance that would have put you near the top of any other section actually lands you at median in your section (and therefore, median in the overall rankings).

No, you don't get how it works.

Let's say out of 10 kids, there's 5 B+'s, 2 A's, 2 B's and 1 C+.

the 2 A's are pacing a 4.0.

Doesn't matter if they're in Section xx or Section yy. I suppose if firms KNEW there was section stacking and a kid writes XXXX Scholarship on the resume, they'd know about the stacked section so wouldn't treat them the same as a student with an identical GPA not in that section. Contrarily, it'd still say Shit School XXX on their website.


But doesn't the fact that the firms would care if they knew imply that GPAs aren't comparable between sections? That's the point where I'm stuck, mentally. If the same quality of work would get you the same GPA (and therefore, the same overall class ranking) no matter which section you're in, then "section stacking" wouldn't really weed out any more scholarship students compared to spreading them evenly across all of the sections.

It seems like section stacking only works if the stacked section is full of overly-bright students compared to the class as a whole. I guess that would be the case for a school that attaches stips to ALL of its scholarships, though.




Return to “Ask a Law Student / Graduate”

Who is online

Users browsing this forum: Chill_Out and 2 guests