Significance of 25th/50th/75th percentile spread?

(Applications Advice, Letters of Recommendation . . . )
User avatar
ScottRiqui
Posts: 3640
Joined: Mon Nov 29, 2010 8:09 pm

Significance of 25th/50th/75th percentile spread?

Postby ScottRiqui » Fri Sep 13, 2013 12:49 pm

I see a lot of schools where the LSAT difference between the 25th and 50th percentiles is about the same as the difference between the 50th and 75th, such as 170/172/174.

But there are other schools where the median is much closer to the 75th percentile, like 164/170/171.

Does the median being so close to the 75th tell us anything beyond just the statistical significance of the numbers? Does it say anything about the school or its admissions policies?

surferkid
Posts: 2
Joined: Thu Aug 22, 2013 2:02 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby surferkid » Sat Sep 14, 2013 1:23 pm

From what I understand it suggests median gaming. They admit a bunch of splitters and reverse splitters, many of whom are just slightly above the schools median on their strong stat, and far below the schools median on their weak stat, in order to get the highest possible median.

User avatar
john1990
Posts: 1211
Joined: Sun Jun 05, 2011 5:49 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby john1990 » Sat Sep 14, 2013 1:38 pm

It may be more difficult to attract students to the school with number far greater than median in that range because they are more likely to attend HYS without $$. There are less student with LSAT's that high with low GPA's as well

User avatar
AT9
Posts: 1305
Joined: Sun May 17, 2009 6:00 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby AT9 » Sat Sep 14, 2013 10:29 pm

surferkid wrote:From what I understand it suggests median gaming. They admit a bunch of splitters and reverse splitters, many of whom are just slightly above the schools median on their strong stat, and far below the schools median on their weak stat, in order to get the highest possible median.


So let's say a school gaming the medians is aiming to maintain their current stats at 168/3.7, and these three applications are under consideration:

A) 168/3.2 (+1, -.5)
B) 164/3.8 (-4, +.1)
C) 166/3.6 (-2, -.1)

Would you expect A and B to have an advantage over C, even though C is closer to the medians overall? I've wondered about this too, and it explains schools like UMN with such high and similar medians/75th's, but comparatively low 25th's.

User avatar
midwest17
Posts: 1686
Joined: Sat Aug 31, 2013 5:27 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby midwest17 » Sat Sep 14, 2013 10:33 pm

AT9 wrote:So let's say a school gaming the medians is aiming to maintain their current stats at 168/3.7, and these three applications are under consideration:

A) 168/3.2 (+1, -.5)
B) 164/3.8 (-4, +.1)
C) 166/3.6 (-2, -.1)

Would you expect A and B to have an advantage over C, even though C is closer to the medians overall? I've wondered about this too, and it explains schools like UMN with such high and similar medians/75th's, but comparatively low 25th's.


I think the more relevant point is that they aren't giving much of an advantage to people who are significantly above their medians compared to people who are just at their medians. So a 172/3.7 might not be any better off than a 168/3.7 (which means they're more likely to end up with the 168/3.7 matriculating, since they're not willing to spend money on the 172.)

User avatar
MoMettaMonk
Posts: 513
Joined: Thu Jul 04, 2013 3:29 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby MoMettaMonk » Sat Sep 14, 2013 10:55 pm

AT9 wrote:
surferkid wrote:From what I understand it suggests median gaming. They admit a bunch of splitters and reverse splitters, many of whom are just slightly above the schools median on their strong stat, and far below the schools median on their weak stat, in order to get the highest possible median.


So let's say a school gaming the medians is aiming to maintain their current stats at 168/3.7, and these three applications are under consideration:

A) 168/3.2 (+1, -.5)
B) 164/3.8 (-4, +.1)
C) 166/3.6 (-2, -.1)

Would you expect A and B to have an advantage over C, even though C is closer to the medians overall? I've wondered about this too, and it explains schools like UMN with such high and similar medians/75th's, but comparatively low 25th's.


If they're really going for the medians, then A and B would definitely have an advantage over C. C might appear to be a more well rounded candidate, but if the main goal is having 50% of your class above two specific numbers, then someone who is below both of those numbers does nothing but hurt you.

splitcity
Posts: 52
Joined: Mon Jul 15, 2013 10:25 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby splitcity » Sat Sep 14, 2013 11:45 pm

midwest17 wrote:
AT9 wrote:So let's say a school gaming the medians is aiming to maintain their current stats at 168/3.7, and these three applications are under consideration:

A) 168/3.2 (+1, -.5)
B) 164/3.8 (-4, +.1)
C) 166/3.6 (-2, -.1)

Would you expect A and B to have an advantage over C, even though C is closer to the medians overall? I've wondered about this too, and it explains schools like UMN with such high and similar medians/75th's, but comparatively low 25th's.


I think the more relevant point is that they aren't giving much of an advantage to people who are significantly above their medians compared to people who are just at their medians. So a 172/3.7 might not be any better off than a 168/3.7 (which means they're more likely to end up with the 168/3.7 matriculating, since they're not willing to spend money on the 172.)



You're assuming that the schools in question base their decisions primarily on the lsat being above median. A school's LSAT median is only 12.5% of it's overall USNWR. Many other factors come into play, such as bar passage rate and performance after law school. A 172 certainly indicates a higher potential for passing the BAR and possibly for doing well at the school and getting a good job after graduation. i'm sure this is also taken into account.

indo
Posts: 288
Joined: Fri Oct 21, 2011 3:02 am

Re: Significance of 25th/50th/75th percentile spread?

Postby indo » Sun Sep 15, 2013 12:21 am

splitcity wrote:
midwest17 wrote:
AT9 wrote:So let's say a school gaming the medians is aiming to maintain their current stats at 168/3.7, and these three applications are under consideration:

A) 168/3.2 (+1, -.5)
B) 164/3.8 (-4, +.1)
C) 166/3.6 (-2, -.1)

Would you expect A and B to have an advantage over C, even though C is closer to the medians overall? I've wondered about this too, and it explains schools like UMN with such high and similar medians/75th's, but comparatively low 25th's.


I think the more relevant point is that they aren't giving much of an advantage to people who are significantly above their medians compared to people who are just at their medians. So a 172/3.7 might not be any better off than a 168/3.7 (which means they're more likely to end up with the 168/3.7 matriculating, since they're not willing to spend money on the 172.)



You're assuming that the schools in question base their decisions primarily on the lsat being above median. A school's LSAT median is only 12.5% of it's overall USNWR. Many other factors come into play, such as bar passage rate and performance after law school. A 172 certainly indicates a higher potential for passing the BAR and possibly for doing well at the school and getting a good job after graduation. i'm sure this is also taken into account.


Are you saying that someone who has got accepted at Yale with LSAT 152 will NOT do well at passing the bar and NOT get good employment ?

http://www.law.yale.edu/admissions/profile.htm
Last edited by indo on Sun Sep 15, 2013 12:45 am, edited 1 time in total.

User avatar
haus
Posts: 2840
Joined: Wed Aug 18, 2010 11:07 am

Re: Significance of 25th/50th/75th percentile spread?

Postby haus » Sun Sep 15, 2013 12:43 am

splitcity wrote:
midwest17 wrote:
AT9 wrote:So let's say a school gaming the medians is aiming to maintain their current stats at 168/3.7, and these three applications are under consideration:

A) 168/3.2 (+1, -.5)
B) 164/3.8 (-4, +.1)
C) 166/3.6 (-2, -.1)

Would you expect A and B to have an advantage over C, even though C is closer to the medians overall? I've wondered about this too, and it explains schools like UMN with such high and similar medians/75th's, but comparatively low 25th's.


I think the more relevant point is that they aren't giving much of an advantage to people who are significantly above their medians compared to people who are just at their medians. So a 172/3.7 might not be any better off than a 168/3.7 (which means they're more likely to end up with the 168/3.7 matriculating, since they're not willing to spend money on the 172.)



You're assuming that the schools in question base their decisions primarily on the lsat being above median. A school's LSAT median is only 12.5% of it's overall USNWR. Many other factors come into play, such as bar passage rate and performance after law school. A 172 certainly indicates a higher potential for passing the BAR and possibly for doing well at the school and getting a good job after graduation. i'm sure this is also taken into account.

lol @ LSAT correlating to anything beyond acceptance at specific law schools.

User avatar
MoMettaMonk
Posts: 513
Joined: Thu Jul 04, 2013 3:29 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby MoMettaMonk » Sun Sep 15, 2013 12:46 am

splitcity wrote:
You're assuming that the schools in question base their decisions primarily on the lsat being above median. A school's LSAT median is only 12.5% of it's overall USNWR. Many other factors come into play, such as bar passage rate and performance after law school. A 172 certainly indicates a higher potential for passing the BAR and possibly for doing well at the school and getting a good job after graduation. i'm sure this is also taken into account.


Actually it doesn't. The LSAT has never been shown to be predictive of anything other than relative success in law school classes, and even then it's a wildly imperfect tool. Besides, within the general score bands that most schools operate (about 7-10 points), the LSAT isn't particularly predictive of anything.

splitcity
Posts: 52
Joined: Mon Jul 15, 2013 10:25 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby splitcity » Sun Sep 15, 2013 12:53 am

I would say that being predictive of "relative success in law school classes" is a pretty damn important thing to be predictive about.




But I guess there's a chance that all of you guys are right and the adcomms, ABA, and rankers are all idiots who irrationally emphasize the importance of the LSAT because they don't have access to the vast database of evidence that you do.

User avatar
haus
Posts: 2840
Joined: Wed Aug 18, 2010 11:07 am

Re: Significance of 25th/50th/75th percentile spread?

Postby haus » Sun Sep 15, 2013 1:11 am

splitcity wrote:I would say that being predictive of "relative success in law school classes" is a pretty damn important thing to be predictive about.

But I guess there's a chance that all of you guys are right and the adcomms, ABA, and rankers are all idiots who irrationally emphasize the importance of the LSAT because they don't have access to the vast database of evidence that you do.

Adcoms are interested in high LSAT scores because USNWR is, not because it is a sign you will crush the bar or become a great lawyer.

I have not seen a lot of love from the ABA for the LSAT.

User avatar
jselson
Posts: 6337
Joined: Sat Jan 05, 2013 3:51 am

Re: Significance of 25th/50th/75th percentile spread?

Postby jselson » Sun Sep 15, 2013 2:48 am

splitcity wrote:I would say that being predictive of "relative success in law school classes" is a pretty damn important thing to be predictive about.




But I guess there's a chance that all of you guys are right and the adcomms, ABA, and rankers are all idiots who irrationally emphasize the importance of the LSAT because they don't have access to the vast database of evidence that you do.


Law schools are interested in the LSAT because 1) USNWR is interested in it, and it's one of the few metrics that a school has significant control over, 2) there are more "plausible" applicants than spots available for acceptance, and many of these applicants are similar overall, so the LSAT provides something that lets the school distinguish plausible applicants, 3) the meaning of gpas amongst individuals is not consistent, 4) the LSAT has some predictive power of success, as an indicator of both the ability to use reasoning that can be helpful on law school exams and the applicant's work ethic, 5) higher LSAT numbers indicate more "prestige" because, the higher they go, the fewer people there are who have them, 6) the ABA requires them to be interested in it to some measure since students applying must take it and many students would feel like their efforts were wasted if it didn't matter, and 7) many of the common arguments students make against the LSAT could easily be made against the other elements of the application (e.g., that a student just "isn't good at writing personal statements," or that gpas are biased against minorities, and such). The sum total of all those things leads to the high value of the LSAT in admissions.

User avatar
john1990
Posts: 1211
Joined: Sun Jun 05, 2011 5:49 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby john1990 » Sun Sep 15, 2013 1:42 pm

Predictive of success in law school is highly relevant to the level of success a law student can expect after law school :roll:

NYstate
Posts: 1566
Joined: Thu Jan 31, 2013 1:44 am

Re: Significance of 25th/50th/75th percentile spread?

Postby NYstate » Sun Sep 15, 2013 1:47 pm

splitcity wrote:I would say that being predictive of "relative success in law school classes" is a pretty damn important thing to be predictive about.




But I guess there's a chance that all of you guys are right and the adcomms, ABA, and rankers are all idiots who irrationally emphasize the importance of the LSAT because they don't have access to the vast database of evidence that you do.


You mean people who have actual experience with the success and failure of their classmates in law school? Yes, we know on a different level how predictive LSAT is on determine success against peers on one exam graded on a strict curve.


While there may be some correlation I when looking at the overall group, the predictive ability within a small law school class doesn't exist, in my experience.

User avatar
midwest17
Posts: 1686
Joined: Sat Aug 31, 2013 5:27 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby midwest17 » Sun Sep 15, 2013 2:15 pm

splitcity wrote:
midwest17 wrote:I think the more relevant point is that they aren't giving much of an advantage to people who are significantly above their medians compared to people who are just at their medians. So a 172/3.7 might not be any better off than a 168/3.7 (which means they're more likely to end up with the 168/3.7 matriculating, since they're not willing to spend money on the 172.)



You're assuming that the schools in question base their decisions primarily on the lsat being above median. A school's LSAT median is only 12.5% of it's overall USNWR. Many other factors come into play, such as bar passage rate and performance after law school. A 172 certainly indicates a higher potential for passing the BAR and possibly for doing well at the school and getting a good job after graduation. i'm sure this is also taken into account.


No, I'm not. I'm saying that when a school has LSAT medians and 75%s that are extremely close together, this is evidence that the school doesn't place much emphasis on attracting people with LSATs significantly above its median.

I'm not saying that schools have no reason to try for outlier LSATs or that no schools do. I'm addressing a very specific case.

User avatar
john1990
Posts: 1211
Joined: Sun Jun 05, 2011 5:49 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby john1990 » Sun Sep 15, 2013 5:28 pm

NYstate wrote:
splitcity wrote:I would say that being predictive of "relative success in law school classes" is a pretty damn important thing to be predictive about.




But I guess there's a chance that all of you guys are right and the adcomms, ABA, and rankers are all idiots who irrationally emphasize the importance of the LSAT because they don't have access to the vast database of evidence that you do.


You mean people who have actual experience with the success and failure of their classmates in law school? Yes, we know on a different level how predictive LSAT is on determine success against peers on one exam graded on a strict curve.


While there may be some correlation I when looking at the overall group, the predictive ability within a small law school class doesn't exist, in my experience.


I would be surprised to find that the students with high LSAT's do not place better in the class than the students with low LSAT's. These students are given scholarships which coincidentally require that they place higher in the class and this is usually the case that they outperform

User avatar
john1990
Posts: 1211
Joined: Sun Jun 05, 2011 5:49 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby john1990 » Sun Sep 15, 2013 5:30 pm

midwest17 wrote:
splitcity wrote:
midwest17 wrote:I think the more relevant point is that they aren't giving much of an advantage to people who are significantly above their medians compared to people who are just at their medians. So a 172/3.7 might not be any better off than a 168/3.7 (which means they're more likely to end up with the 168/3.7 matriculating, since they're not willing to spend money on the 172.)



You're assuming that the schools in question base their decisions primarily on the lsat being above median. A school's LSAT median is only 12.5% of it's overall USNWR. Many other factors come into play, such as bar passage rate and performance after law school. A 172 certainly indicates a higher potential for passing the BAR and possibly for doing well at the school and getting a good job after graduation. i'm sure this is also taken into account.


No, I'm not. I'm saying that when a school has LSAT medians and 75%s that are extremely close together, this is evidence that the school doesn't place much emphasis on attracting people with LSATs significantly above its median.

I'm not saying that schools have no reason to try for outlier LSATs or that no schools do. I'm addressing a very specific case.


I agree that there is much more emphasis on building the schools medians. While the 75th percentile is important it is not as crucial. It is also harder to build

splitcity
Posts: 52
Joined: Mon Jul 15, 2013 10:25 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby splitcity » Sun Sep 15, 2013 8:31 pm

The LSAT is designed by professionals specifically to test your aptitude for law school (hence Law School Aptitude Test). It's kind of ridiculous to argue that it in no way indicates how you will do in law school. If you get a 146 or something and you don't have any knowledge of the basics of logic or you can barely get through 2 RC passages because your reading comprehension skills aren't that great, then its safe to say your odds of doing well at an elite school are slim.

That's not to say there aren't exceptions. Law school is still a school and its purpose is to develop your skills so it is possible that someone with a 150 can somehow get into a good school and hone their skills to the point where they become a good lawyer.

BTW

I think getting 75th percentile LSATs has its advantages too. The more data points you have above your median, the higher your median gets.

User avatar
MoMettaMonk
Posts: 513
Joined: Thu Jul 04, 2013 3:29 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby MoMettaMonk » Sun Sep 15, 2013 8:45 pm

splitcity wrote:The LSAT is designed by professionals specifically to test your aptitude for law school (hence Law School Aptitude Test). It's kind of ridiculous to argue that it in no way indicates how you will do in law school. If you get a 146 or something and you don't have any knowledge of the basics of logic or you can barely get through 2 RC passages because your reading comprehension skills aren't that great, then its safe to say your odds of doing well at an elite school are slim.

That's not to say there aren't exceptions. Law school is still a school and its purpose is to develop your skills so it is possible that someone with a 150 can somehow get into a good school and hone their skills to the point where they become a good lawyer.

BTW

I think getting 75th percentile LSATs has its advantages too. The more data points you have above your median, the higher your median gets.


http://www.youtube.com/watch?v=_7_xHsce57c Go to about the 15:00 mark and watch his explanation of how the LSAT correlates to law school performance.

Can you explain what you mean by the bolded?

User avatar
midwest17
Posts: 1686
Joined: Sat Aug 31, 2013 5:27 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby midwest17 » Sun Sep 15, 2013 9:48 pm

MoMettaMonk wrote:http://www.youtube.com/watch?v=_7_xHsce57c Go to about the 15:00 mark and watch his explanation of how the LSAT correlates to law school performance.


Thanks, that was interesting. That's pretty remarkable, that a standardized test is actually that predictive of performance.

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

Re: Significance of 25th/50th/75th percentile spread?

Postby ScottRiqui » Sun Sep 15, 2013 9:52 pm

midwest17 wrote:
MoMettaMonk wrote:http://www.youtube.com/watch?v=_7_xHsce57c Go to about the 15:00 mark and watch his explanation of how the LSAT correlates to law school performance.


Thanks, that was interesting. That's pretty remarkable, that a standardized test is actually that predictive of performance.


I haven't watched the video yet since Breaking Bad is on, but is it the "LSAT plus uGPA correlates to 1L performance with a correlation factor of 0.4"?

User avatar
midwest17
Posts: 1686
Joined: Sat Aug 31, 2013 5:27 pm

Re: Significance of 25th/50th/75th percentile spread?

Postby midwest17 » Sun Sep 15, 2013 9:54 pm

ScottRiqui wrote:I haven't watched the video yet since Breaking Bad is on, but is it the "LSAT plus uGPA correlates to 1L performance with a correlation factor of 0.4"?


I think he was saying that LSAT alone has a correlation factor of 0.4, and adding uGPA bumps it up a little past that.




Return to “Law School Admissions Forum”

Who is online

Users browsing this forum: Baidu [Spider], Bing [Bot] and 5 guests