Law School Discussion

LSAT correlation to Law school performance

#2

LSAT correlation to Law school performance
« on: November 28, 2005, 11:58:08 PM »
How do you guys think one's LSAT score correlates to law school performance?  To give a specific example do you think someone with a LSAT score near that school's 75th% would be much more likely to do better than a student at the same school with an LSAT at the 25%? 

Also, when thinking about this question I'm curious to see what you think of how one's practice scores factor into this.  For example, is someone had been practicing consistently at, say, 177 and choked on the test day and scored a 165 and went to, say, BU do you think that person would be at an advantage intellectually in terms of their classmates?  Just something to ponder.

thescreed

Re: LSAT correlation to Law school performance
« Reply #1 on: November 29, 2005, 12:30:00 AM »
If someone was practicing consistently at 177 and actually scored 165, I'd say that person was somehow cheating during practice.

Slow Blues

Re: LSAT correlation to Law school performance
« Reply #2 on: November 29, 2005, 06:20:49 AM »
How do you guys think one's LSAT score correlates to law school performance?  To give a specific example do you think someone with a LSAT score near that school's 75th% would be much more likely to do better than a student at the same school with an LSAT at the 25%? 

Also, when thinking about this question I'm curious to see what you think of how one's practice scores factor into this.  For example, is someone had been practicing consistently at, say, 177 and choked on the test day and scored a 165 and went to, say, BU do you think that person would be at an advantage intellectually in terms of their classmates?  Just something to ponder.

Take a look at the first three paragraphs on page 18 of the below.

http://www.bcgsearch.com/crc/book2005/bcg_guide_2005.pdf

Quote
Finally, what about the law student at one of the lesser highly competitive law schools? Is a student
who finishes in the top 10% of his/her class at, say, Duke or UCLA the equal of a Harvard
student who only finished in the top half of his/her class?

Justin N. Bezis specializes in intellectual capital and the American society’s strategies for its
deployment. In a recent article “An Inquiry into the Implications of Using Percentage Rankings of
Heterogeneous Scholastic Populations,” Bezis asks how, if comparing two types of law schools—a
prestigious national one and a less-prestigious regional one—a law firm can “(1) maximize the
chance of selecting a high-skill student and (2) minimize the chance of selecting a low-skill student.”
This sounds like a simple question, but law firms have differing ways of dealing with this
question, says Bezis. One way, Bezis mentions, is for a law firm to interview “about the top 5 to
10% at the regional law school and the top 10 to 20% at the national law school. The logic of this
approach is based on the assumption that the top 20% of the national law school is likely to be
similar in “quality” to the top 10% at the regional law school. This is where Bezis’ research gets
interesting.

Bezis assumes for purposes of discussion that both the national law school and the regional law
school use just two grades: Exemplary and Pass. He goes on to postulate that “both schools base
class rank on the number of Exemplary grades that each student received.” He further assumes
that both schools use five evaluation events to determine class rank. “Students in the first strata
have all E grades, students in the second strata have four E grades, and students in the third strata
have three E grades, etc.”11 Now let’s assume that the law school populations of both the national
and the regional law schools are broken into high, medium, and moderate levels, based on college
GPA and LSAT scores.

Bezis’ research finds that “…high-level students have an 80% predictability of an E grade in each
course and a 20% change of a Pass grade in each course. Medium-level students have a 70%
chance of an E and a 30% chance of a Pass grade. Moderate-level students have a 50% chance of
an E and a 50% chance of a Pass grade.”

Such predictions fall within the realm of the assumed and for this reason may not seem particularly
newsworthy. Bezis notes on page two of his study that one would expect to find “the cream
of the crop” rising to fill the top stratum. That would be what conventional wisdom would predict.
Instead, what Bezis actually found was that “less than 15% of students in the highest stratum had
the highest LSAT and college GPA scores.”12

What we learn from the Bezis study is that law firms’ conventional wisdom is both flawed and not
flawed. It is flawed in its assumption that students in the top 10% at an elite school are superior
to other students in that school. Statistically, Bezis says, the top 10% of any class will not be filled
with only the students with the highest LSAT and college GPA scores. When recruiting at an elite
school, if you recruit only from the top 10%, you may be missing superior law firm candidate who
ranked lower in terms of pure academics.

Hybrid Vigor

  • ****
  • 1924
  • prestigious
    • View Profile
Re: LSAT correlation to Law school performance
« Reply #3 on: November 29, 2005, 08:04:03 AM »
Considering all I've heard about law school grades is that they are NOT a measure of how well you know the law, but rather how well you can respond to a given professor's questions, I'd say no. I know people who for a fact had GPAs and LSAT in the bottom 25% of their incoming class, but graduated in the top third. Of course, this ancedotal evidence doesn't disprove the idea that LSAT can predict grades. Is a person in the top 25% LIKELY to do better? Of course they are. But that's an aggregate...you can't break it down to saying any one person can't do well b/c of their LSAT/GPA.

Typhoon Longwang

  • ****
  • 719
  • I have the power to stop an LSAT!
    • View Profile
Re: LSAT correlation to Law school performance
« Reply #4 on: November 29, 2005, 08:12:01 AM »
Take a look at the first three paragraphs on page 18 of the below.

http://www.bcgsearch.com/crc/book2005/bcg_guide_2005.pdf

Great guide, thanks for posting it.


bulk


RocketBot

Re: LSAT correlation to Law school performance
« Reply #7 on: November 29, 2005, 09:09:28 AM »
I remember reading that it's the best predictor for first year grades -- but even then, it's only around 20% accurate

I hate how many people use this number incorrectly.  The fact of the matter is that it is just wrong to say that only 20% of students' first year grades are accurately predicted by the LSAT.

Thought experiment: What would it even mean for a student's LSAT to accurately predict 1L grades? Obviously, if a kid at a school's 75th% LSAT scored in the 75th% of his class it would be an accurate prediction, but what if he/she scored in the 76th%? 80th%? 85th%?  We don't even have a reference point for what is considered "accurate," yet people throw around the number like they were on the research team that found it.

The number itself was taken from a poorly-wri Wikipedia entry and references a .2 correlation coefficint, which says nothing about how many students have their performance accurately predicted by their LSAT score (if we even had a definition for accurate), but only speaks to how much the LSAT reducess the overall randomness of 1L grades (which, statistically speaking, is quite a lot).

Exactly.  And those who use that Michigan number only do so to suit their purposes (or they have no idea how to do research outside of Wikipedia).  Here is the relevant quote from LSAC:

During 2001, validity studies were conducted for 181 law
schools. Correlations between LSAT scores and first-year law
school grades ranged from .09 to .58 (median is .41).
Correlations between LSAT scores combined with
undergraduate grade-point averages and first-year law school
grades ranged from .28 to .66 (median is .49).


0.41 Median.  So stop throwing around that 0.2 or "20%" number.

#2

Re: LSAT correlation to Law school performance
« Reply #8 on: November 29, 2005, 09:17:10 AM »
http://powerscore.com/lsat/help/salary.htm


I don't think this is really saying much because IMO, these salary differentials are based much more on the reputation and prestige of the school than they are about the LSAT.  It also doesn't say much about the individual.  For instance, even if someone did have a 180 from Cooley and finished #1 in their class they would have extreme difficulty getting the same job of someone who finished #1 at Georgetown.

bulk

Re: LSAT correlation to Law school performance
« Reply #9 on: November 29, 2005, 09:46:31 AM »
http://powerscore.com/lsat/help/salary.htm


I don't think this is really saying much because IMO, these salary differentials are based much more on the reputation and prestige of the school than they are about the LSAT.  It also doesn't say much about the individual.  For instance, even if someone did have a 180 from Cooley and finished #1 in their class they would have extreme difficulty getting the same job of someone who finished #1 at Georgetown.

lsat > school > job > salary

correlation ^^^