Every question on every version of the test is graded and analyzed w/ statistics, to get a mean, median, mode and standard deviation. Based on how well everyone does on one particular question, it is given a certain amount of "points". So, for each question on the test that you take, they do the statistics, and then apply them to the 15 point scale for each section. Because they are grading on a statistical analysis of each question, it comes out to a curve, as opposed to strictly number correct divided by number incorrect. That is why it takes so long to grade the MCAT, in addition to graders actually reading your essays.

I hope that helps.