ChampagnePapi wrote:UtilityMonster wrote:bk187 wrote:Hutz_and_Goodman wrote:Given the OPs numbers, he could attend a T20 in a full ride. It would be insane to pay sticker at NYU.
I'm not saying that NYU at sticker is sane, but these two things are not really comparable. The types of jobs that a majority of the students at GW/WUSTL get look nothing like the types of jobs that a majority of the students at NYU get.
Of course this isn't all that important for OP since OP has numbers for significant T14 schollies (if not full rides), I'm merely nitpicking your T20 comment.
As you should, which is why "T20" should never be used to group law schools.
bk187 wrote:Hutz_and_Goodman wrote:Someone with his numbers will be top quarter at GW, minimum. Probably top 10/15%.
lolno
I am curious how strong the r^2 is for LSAT/GPA numbers and performance in law school. It is pertinent to a number of questions people pose on these boards, but not one seems to know it. If we knew it, the advice we could offer would be so much more precise.
Its around 0.4 IIRC. I wonder if they've done a study with only T20's. If I had to guess I bet the correlation is weaker, but I have no foundation for that
This may be the r^2, but a major issue, especially at a lot of top schools, is restriction of range. The way top schools admit people, the top half of the class is clumped around median or very slightly above. Thus, even something with an r^2 of 0.4 (which means it really only explains 16% of your GPA variance), the predictability of your LSAT as to whether you end up median or at the top of the class is probably pretty close to nothing. Their predictability over the bottom half of the class is probably a lot higher, since there's a much larger range there. This is even more true if the school is splitter-friendly and has 25/50/75 LSAT median scores of like 169/168/162. LSAT/uGPA might be good enough to say "it is likely that you won't end up at in the bottom 25% if you're above the class's 25th percentile in each," but predicting where in the top 50-75% of the class you will end up is pretty much a toss up.
This may be different if you are
way above the school's entering class statistics (for example, if you have a 170/3.9 and go to Brooklyn instead of NYU), but simply going to George Washington with a 170 on your LSAT when their median is 167 means almost nothing. Also, keep in mind that LSAC reports the margin of error of your LSAT score to be 2.7 points. So really, a 170 isn't different than a 167 to some level of statistical certainty.
The problem is that law school exams are not standardized or professionally developed. There's very little quality control, and they are highly unreliable and only marginally valid at measuring any sort of skills that predict any sort of outcome. Because the error variance in law exam scores is so high, and reliability is a major issue, there's simply not a lot of factors that can predict your performance on them to any significant degree.