The raw scores for the physical science, verbal reasoning and biological sciences sections are derived from the number of questions that a test taker answers correctly out of 52, 40 and 52 respectively. These raw scores are converted to scaled scores of 1-15 based on a formula that varies from test to test. The formula also varies by section; for instance, missing five questions on the verbal reasoning section will result in a significantly lower scaled score than missing five questions on the physical science or biological science sections. The scaled scores are added together to form a score out of 45.
The writing sample is scored based on grades assigned to each of two essays by both a human and a computer. The human and the computer grade both essays, assigning a score of between 1 and 6 to each essay. The scores for the two essays are then added together to produce a cumulative score between 2 and 12. This score is converted to an alphabetical scale ranging from J, representing 2, to T, representing 12.
Traditionally, the score on the writing sample is considered relatively unimportant; in many cases, only a very high score (S or T) or a very low score (J or K) will affect a student's admission prospects. Some medical schools, however, have begun to install absolute cut-off scores for the writing sample. Different schools are known to favor different multiple-choice sections, but consistency on these sections is generally preferred to lopsided scoring.
The majority of competitive medical schools prefer a numerical score of at least 30 and a writing sample score of at least P. In 2008, the average score for successful U.S. medical school applicants was 30.9P.
Certain questions on every MCAT are not scored. These are either included as "testing questions" for future MCATs or discarded if they are later deemed too ambiguous or difficult. Contrary to popular belief, it is not easier to score higher during the winter session, when weaker students are understood to take the exam. The providers of the test have successfully ensured that there are no consistent scoring discrepancies between test dates.