Re: Iowa with full-ride, Northwestern/$10k,or michigan sticker?
Posted: Mon Jun 03, 2013 1:38 am
Law School Discussion Forums
Monochromatic Oeuvre wrote:NYstate wrote: Are you an 0L? I'm trying to decide if the only people who cite that correlation as actually meaningful are the ones who have never taken a law school exam. So far, that seems to be the case, but I could be wrong. My theory is that reliance on statistical correlation breaks down after actual, individual experience.
I don't understand what you're saying. The correlation is by definition meaningful (given that it's statistically significant) because it identifies an association pertaining to important results and demonstrates at least some predictive power with regard to those results. Correlation isn't causation, but the suggestion is strengthened by a number of factors: a) causality can't possibly run in the other direction (A can only cause B, and not vice versa), b there's no reason to believe predictive ability is different in the near future than in the recent past, and c) most importantly, there is a strong logical connection between the factors. The causal link is third-cause indirect--those who are smart and work hard get good GPAs/LSATs, and those who are smart and work hard also get good 1L grades--but certainly makes sense.
There is an important factor to distinguish--intraschool vs. interschool. If you have the best incoming numbers for anyone at CCN, for example, your combined index number (aka combined LSAT/GPA) is probably very close to all your classmates' index numbers. Therefore even if you have a decent correlation, you're not a lock to finish above median, because everyone else has credentials that are pretty close to yours (given that they all got into the same school you did). However, given that same set of numbers, if you take them to a TTT, however, you will almost certainly finish above median, as you are significantly more capable than the rest of your classmates.
If you presume a correlation of .5 for the index number (and that it's uniformly linear), then the absolute best kid in the incoming class is still only 75% to finish above median--not a guarantee by any means. The kid is also barely more than a coin flip to finish in the top 30%. These aren't numbers that guarantee a job unless you're HYS bound (and I think we can agree that if you're the best incoming kid at HYS, you don't need practical advice).
So it matters, but the variability is greater than firm variability in selecting from a candidate, so it's not a factor that your decision is going to hinge on. By the time you get down to a school where you have that predictive validity, the firm cutoff point is higher than where you would be predicted to finish. Ex. if you'd be 50% to finish top half at Chicago (an average matriculant), then it doesn't make to sense to go be the best student at Minnesota (~75% chance at top half). If firm A hires top half at Chicago and top 5% at Minnesota, then your chances of getting that job are 50% at Chicago and a little under 10% from Minnesota.
TL;DR--The correlation is a factor, but it should never override placement and money.