Is there really a correlation? I find it somewhat implausible, because say you got a really bad score on the SAT, and you prep for the LSAT for a year, you're telling me that person won't get a good score because the SAT was low? I really don't understand how there can be a correlation when you can do well on either one by preparing adequately. There has to be people out there who have a low SAT, high LSAT or vice versa.
Composition fallacy. You can easily have outliers in your regression but maintain a valid correlation. In layman's terms, the correlation has no effect on an individual's scores, but when we look at all the LSAT takers as a whole, we see their LSAT scores generally correlate to their SAT scores.