There is no test in America that has quite the reputation of the Scholastic Aptitude Test, or SAT. Colleges use it to determine who gets in and who does not, and (a disturbing number of) teens think it measures how intelligent they are. Companies charge hundreds of dollars for a hundred points on it, and an entire industry is built on providing teens with study materials.
So. What do the numbers actually mean? The answer is profoundly simple, but at the same time it contains a subtle complexity because of the statistics used to produce it.
Let’s start with the easiest bit. The SAT, in its current form, has ten sections that the test taker fills in with bubbles. Of these, one is an “experimental” section, used to evaluate questions for possible future tests. The other nine are from the SAT’s three big categories: Critical Reading, Mathematics, and Writing. Writing is the newest section, and part of it consists of an essay. Each of the three sections is scored from 200 to 800, meaning that the entire test is scored from a measly 600 to the legendary 2400.
The score for each section is based on a “raw score,” which is the student’s total number of correct answers on all sections in a category, minus the number of wrong answers * ¼. When I talk later about the “average” score on a section, I am referring to this average raw score. It should be noted that all questions are weighted equally, regardless of difficulty. This means that a right answer to a hard question is considered the same as a right answer to an easy question.
As a general rule, the average score for each section will be assigned a point of value of 500, which also happens to be the average of the highest and lowest possible score on each section. Thus, the 50th percentile is 500 on a section, and 1500 on the test as a whole. To determine the rest of the scores, a tool called the standard deviation is used. Although the details of the standard deviation can be a bit hard to understand, suffice to say that the “first standard deviation” is at about the 66th percentile (and at the corresponding spot on the other side of 50, i.e. the 34th percentile), the “second standard deviation” is at the 95th percentile (and the 5th), and the third is at the 99th (and 1st). The scores are then distributed in such a way that a person whose raw score is at the 66th percentile will be as far on the final score out of 800 will be as far from the 50th percentile, as he himself is from the 95th.