Data from medical school web sites was compiled from 2000 - 2003 in 5 sets of data. Missing data was filled in with adjacent year data. The Bioscience, Physical Science, and Verbal Reasoning section scores were averaged for each school. This was averaged for the 2000 - 2003 scores to gain a final average. This is compared to the 2005 match data on choice of family medicine as published in Family Medicine.
MCAT Central: Changes, Impacts on Distribution and Career Choice
Regressions confirm the relationship. Choice of Family Medicine Regression
Also other graphics comparing MCAT scores for those that Drive Difficulty or Distinction reveal the folly of admissions stacked to the right since there is little risk for the usual student. The flat curve means that students can have high, medium, or low levels of performance from similar starting points. Admitting exclusive students in terms of scores mainly excludes the students most likely to choose locations outside of major medical centers, family physicians.
Admissions for research potential also makes little sense. Research By the Ages
Best Urban Underserved Distributions - lower scoring schools score higher in distribution
Best Rural Underserved Distributions