SAT Score Calculator

How the digital SAT is scored

The SAT reports two section scores and one total. Reading and Writing is scored from 200 to 800, Math is scored from 200 to 800, and the total is the two added together, so it runs from 400 to 1600. There is no separate essay score any more and no penalty for a wrong answer, which means a blank is worth exactly as much as a wrong guess. If you can eliminate one choice, guess.

The test itself is shorter than the paper version most parents remember. Reading and Writing is 54 questions across two modules of 27, and Math is 44 questions across two modules of 22, for 98 questions in about 2 hours and 14 minutes. Each section is adaptive: the second module is chosen based on how you did on the first, so two students can sit the same test date and see questions of different difficulty.

Because of that adaptivity, the same number of correct answers can produce different scaled scores on different forms. A 44 out of 54 on one test date might map to 690 and on another to 670. The scoring is equated so that a 690 means the same thing from year to year, but the raw score behind it is not fixed. That is the single most important thing to understand before you read any conversion table, including the one on this page.

Raw scores to scaled scores: why no official table exists

College Board does not publish a raw to scaled conversion table for the digital SAT. It publishes the scale, the score range and the percentile ranks, and it keeps the equating process internal. Every third party table you see, including this one, is an estimate built by fitting a line to the two endpoints and the midpoint of the scale.

The fit is simple. A raw score of zero maps to 200, a perfect raw score maps to 800, and the points in between are spread evenly. For Reading and Writing that means each of the 54 questions is worth about 11.1 scaled points. For Math, each of the 44 questions is worth about 13.6 points, which is why a single missed Math question costs more than a single missed reading question.

Real tests are not perfectly linear. The middle of the curve is flatter than the ends, because a question near the middle of the difficulty range separates students better than one at either extreme. In practice that means the estimate below is closest at the top and bottom of the range and can be off by 20 to 40 points in the middle. Use it to see where you stand and to set a target, not to predict a score report.

Turning raw answers into section scores

The calculator above uses the linear estimate. Enter the number you got right in each section, out of 54 and 44, and it returns an estimated section score rounded to the nearest 10 and an estimated total. The table shows the same math at a few common raw scores so you can sanity check the result.

Read the table by column, not by row. The left pair and the right pair are separate scales, and a raw score of 40 means something different in each. Someone at 40 of 54 in Reading and Writing and 35 of 44 in Math lands at 640 plus 680, or 1,320. Someone at 45 and 30 lands at 700 plus 610, or 1,310. Nearly the same total, built from opposite strengths.

That is why the section split matters as much as the total. A 1,320 with a 700 reading score and a 620 math score opens different doors than a 1,320 with the numbers reversed, especially for engineering or business programs that look at the math section on its own.

Superscoring: taking the best of each section

A superscore combines your highest Reading and Writing score from any sitting with your highest Math score from any sitting. Sit three times and score 620 and 650, then 680 and 690, then 660 and 700, and your superscore is 680 plus 700, or 1,380. Your best single sitting was 1,370, so the superscore adds 10 points without a single new question answered.

Most colleges accept superscores, and many calculate one automatically from every score you send. A few do not, and a few ask you to report every sitting. The policies live on each admission office's testing page, and they change, so check the schools on your list before you decide how many times to test.

Superscoring rewards a retake even when the total does not move. If your math score jumps 40 points while reading drops 30, the single sitting total rises by 10 but the superscore rises by 40. That is the argument for retaking after a lopsided result rather than a flat one: a section specific gain is never wasted.

One limit worth knowing: a superscore cannot fix a weak section. If math sits at 560 across three attempts, the superscore still shows 560. Time spent on the weaker section pays more than another attempt at the stronger one.

Percentiles move every year

A percentile tells you what share of test takers scored below you, and it is tied to a specific group in a specific year. College Board recalculates percentiles every year using the scores of that year's graduating class, so the same 1,200 can be the 74th percentile one year and the 72nd the next. Comparing your percentile to a friend's from three years ago is comparing two different rulers.

The averages behind that table move too. The mean total score sits near 1,050 in recent graduating classes, and the number of test takers swings by hundreds of thousands from year to year, which shifts every percentile. A score that was 90th percentile in a small cohort year can slip a point or two when a larger, stronger cohort tests.

For school comparisons, use the middle 50 percent range that each college publishes rather than a national percentile. A school that reports a middle 50 percent of 1,350 to 1,500 is telling you that a quarter of its admitted class scored below 1,350, which is a more useful signal than a national ranking. The federal College Scorecard collects those ranges in one place and lets you sort schools by the numbers rather than by reputation.

What score do you need for a given school

There is no universal cutoff. Colleges publish a middle 50 percent range for admitted students, and landing inside that range puts you in the conversation without guaranteeing anything. The ranges below are typical shapes rather than specific schools, and every number should be checked against the admission office page for the year you apply.

• Most selective private universities: middle 50 percent around 1,470 to 1,560. A 1,500 is competitive, a 1,400 needs a strong rest of file.
• Selective state flagships: around 1,250 to 1,420 for in state applicants, higher for out of state at some campuses.
• Regional public universities: around 1,050 to 1,220, with many admitting well below the range.
• Open admission community colleges: no SAT requirement at all, placement is handled by their own math and English tests.

Test optional policies complicate the picture. A school that does not require a score will still read one if you send it, and a score below the middle 50 percent range can hurt more than no score at all. The rule of thumb admissions officers repeat: send a score at or above the 25th percentile of the previous class, and hold it back if it sits below.

Majors matter at the margins. Engineering and computer science programs often look at the math section with more weight than the total, and some honors colleges publish a separate minimum. A 1,320 with a 720 math score reads differently to an engineering committee than a 1,320 with a 600.

Retaking, Score Choice and what to do next

Most students improve on a retake, and the gain is largest between the first and second attempt, when familiarity with the format does the work. A typical second attempt moves 30 to 60 points. A third attempt usually moves less, which is why two well prepared sittings beat four casual ones.

Score Choice lets you decide which sittings to send, school by school, and it is free. Some colleges still ask for every score, and a few scholarships use all sittings rather than a superscore, so read the policy before you hide a date. Sending a lower score alongside a higher one rarely hurts at a superscore school, because the school builds its own combination.

Preparation that moves the number is boring and specific. Take one full practice test under real timing, mark every question you missed, and sort the misses into three buckets: did not know the content, knew it but ran out of time, and misread the question. Content gaps are fixed with review. Timing is fixed with pacing drills, roughly 71 seconds per reading question and 95 seconds per math question. Misreads are fixed with a marking habit, circling the actual question before you look at the answer choices.

Register through the official SAT registration page, and check the score understanding guide when a report arrives, since it explains the benchmarks and the skills breakdown. National education statistics, including average scores by state, sit in the NCES Digest of Education Statistics.

Scores, percentiles and admission ranges change every year. Confirm the current figures with College Board and with each college before you rely on them.