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).