Something recently occured to me. There is a realy sense in which admissions data can be misleading. It is tempting to think that 25th and 75th percentile LSAT scores indicate the make-up of an entering class. However, unless you are Yale, the majority of the top end are probably only applying as a safety. So you are not really competing against those people for spots in the class, because they are unlikely to take safety schools up on offers unless things go poorly elsewhere.
Think about this: Annually, there are less than 1,500 people who are in the top 1% of LSAT takers. (Less than 150,000 people take the test per year.) The top five schools metriculate well over 1500 students every year. The same people that got accepted to these schools applied to a lot of other schools in the T14, and most are not going to those other t14 schools.
This leads me to believe that the sort of people you are actually competing with for spots (those that might actually take your seat) have LSATs of a lower medium than the data might lead one to believe.
(This post engages in a number of simplifications--such as the assumption that people go to the highest ranked school to which they get accepted--but I think the main point remains valid.)