Law School Discussion

LSAT correlation to Law school performance

sck

  • ****
  • 908
    • View Profile
Re: LSAT correlation to Law school performance
« Reply #20 on: November 29, 2005, 02:12:28 PM »
Man, this conversation is making me want to go pick up my stats books again and refresh my memory.

I hated stats until I got into graduate school, then all of a sudden... it made sense!

that happens to 1 in 3 statistics students.

*snerks*

Funny, that happened to be the number of times I took stats until I liked it....

1st time: failed
2nd time: passed but hated (same class as the fail)
3rd time: Loved it (in grad school)

TheGunner

  • ****
  • 107
  • Seems simple enough...
    • View Profile
Re: LSAT correlation to Law school performance
« Reply #21 on: November 29, 2005, 11:10:18 PM »
At the risk of turning this into a statistics lecture, wouldn't it be more useful to regress LS GPA against LSAT score and take the adjusted R squared value to find exactly how much of GPA can be explained by the LSAT score?  Our hypothesis is that the LSAT will have some predictive value on LS GPA, not simply that LSAT and LS GPA are correlated. 

I'm not sure where this "correlation" statistic came from but it seems to me that a regression would be a better way to determine the impact of LSAT on LS GPA.

Moreover, I don't know how much I trust one score on the LSAT.  I believe the LSAT to have a pretty wide interval of potential scores for a given "talent level", particularly given the discrete nature of the logic game section (either you "get" the game or you don't).

bulk

Re: LSAT correlation to Law school performance
« Reply #22 on: November 29, 2005, 11:15:41 PM »


buelleeer? 

Steve Butabi

  • ****
  • 748
  • and I was like "EMILIOOOO!"
    • View Profile
Re: LSAT correlation to Law school performance
« Reply #23 on: November 30, 2005, 07:52:00 AM »
this debate is pointless and is never used for anything besides lower-scoring people to feel better about themselves or high scoring people to stroke their own ego.


didnt read the thread, jsia

RocketBot

Re: LSAT correlation to Law school performance
« Reply #24 on: November 30, 2005, 11:45:02 AM »
At the risk of turning this into a statistics lecture, wouldn't it be more useful to regress LS GPA against LSAT score and take the adjusted R squared value to find exactly how much of GPA can be explained by the LSAT score?  Our hypothesis is that the LSAT will have some predictive value on LS GPA, not simply that LSAT and LS GPA are correlated. 

I'm not sure where this "correlation" statistic came from but it seems to me that a regression would be a better way to determine the impact of LSAT on LS GPA.

Moreover, I don't know how much I trust one score on the LSAT.  I believe the LSAT to have a pretty wide interval of potential scores for a given "talent level", particularly given the discrete nature of the logic game section (either you "get" the game or you don't).

The correlation coefficient in question (0.41 or 0.2 depending on who is asked)

It's not depending on who is asked.  It's 0.41.  The 0.2 number is specific to Michigan, and should not be used on the whole.  0.41 is the median for all schools in the most recent, official study.

thescreed

Re: LSAT correlation to Law school performance
« Reply #25 on: November 30, 2005, 12:11:22 PM »
this debate is pointless and is never used for anything besides lower-scoring people to feel better about themselves or high scoring people to stroke their own ego.


didnt read the thread, jsia

The reason this debate is pointless is that a large number of people refuse to let facts contradict their intuition about how the world works and/or their opinion about how the world should work. Not coincidentally, this is how George W. Bush operates -- it's characteristic of mediocrity.