SAT Score Chart: Raw Score to Scaled Score

College Board equates every test form so that a 1400 in October means the same thing as a 1400 in March, even though the forms differ in difficulty. Equating is done after the fact, per form, which is exactly why no single conversion table exists to publish.

On top of that, the digital SAT is adaptive. Your scaled score depends not only on how many questions you got right but on which Module 2 you were routed into and how hard the questions you answered were. Two students with the same raw count can end up with different scaled scores.

So treat every score chart you find, including this one, as an estimate built from reported data points rather than an official table.

Out of 54 Reading and Writing questions, these are the approximate anchor points our score calculator uses: 54 correct maps to about 800, 50 to about 760, 46 to about 700, 40 to about 640, 34 to about 580, 27 to about 510, 20 to about 440, 10 to about 340, and 0 to 200.

Between those anchors the relationship is roughly linear. Notice how flat the top is. Going from 46 correct to 50 correct is worth about 60 points, while going from 20 to 27 is worth about 70 points for the same seven question gain. The curve is steepest in the middle.

Out of 44 Math questions, the approximate anchors are: 44 correct maps to about 800, 40 to about 760, 36 to about 700, 31 to about 640, 26 to about 580, 22 to about 530, 16 to about 450, 8 to about 330, and 0 to 200.

Math has fewer questions than Reading and Writing over the same 600 point range, so each Math question is worth more on average. That is one reason a single careless Math mistake tends to cost more than a single Reading and Writing mistake.

Use it for direction, not for precision. If your practice raw scores are climbing, the estimated scaled score climbing is real information. If your estimate says 1310 and you need 1350, that gap is inside the margin of error and does not tell you much.

The one place estimates are genuinely useful is planning. If you know a section score of 700 takes roughly 46 of 54 Reading and Writing questions, you know you have about eight to spare, which changes how you triage under the timer.

Remember that 8 of the 98 questions are unscored pretest items, 2 per module. When you score a practice test, you generally cannot separate them out, so a raw count out of 54 and 44 is the working approach even though only 50 and 40 actually count.

This is one more reason the conversion is fuzzy at the edges. If two of the questions you missed happened to be pretest items, your real raw score was higher than the one you calculated.