How SAT Scores Are Built — The Curve, the Scale, and the Raw Numbers

You get your SAT score back and it says 1180. Or 1340. Or 990. That number feels solid, final, like it was carved into stone the moment you filled in your last bubble. But that number went through a machine before it reached you — a conversion process that most students never learn about and most adults can't explain. Knowing how the sausage gets made won't change your score, but it will change how you think about it.

The Reality

Your SAT score starts as a raw score. That's simply the number of questions you got right. There's no penalty for wrong answers on the current SAT — that changed in 2016 — so your raw score on each section is just correct answers counted up. The Evidence-Based Reading and Writing section (EBRW) combines a Reading Test (52 questions) and a Writing and Language Test (44 questions). The Math section has 58 questions total, split between a no-calculator and a calculator portion. [VERIFY section question counts for current digital SAT format]

Those raw scores then get converted to scaled scores through a process the College Board calls "equating." This is what people loosely refer to as "the curve," but it's not a curve in the way your chemistry teacher uses that word. Your chemistry teacher might add 10 points to everyone's grade because the test was too hard. SAT equating is different. It adjusts scores so that a 600 on one test date means the same thing as a 600 on a different test date, even if one version of the test was slightly harder or easier. The College Board uses statistical methods and data from pretested questions to calibrate this. The goal is consistency across administrations, not generosity.

This means two students who answered the same number of questions correctly on different test dates might receive slightly different scaled scores. On an easier form, you might need 50 out of 52 correct on the Reading Test to hit a certain scaled score. On a harder form, you might get there with 48. The conversion tables are different for every single test form, and they're not published in advance. The College Board releases some of them after the fact in their "Understanding SAT Scores" documents and in QAS (Question-and-Answer Service) reports, but you'll never know your specific conversion table before you sit down on test day.

The Play

The scaled scores land on a range of 200-800 per section, giving you a total between 400 and 1600. The midpoint of that scale is 1000, which is worth keeping in mind — it means the scale is designed so that "average" falls right around the center. According to the College Board's Total Group Profile, the actual mean score for recent graduating classes has hovered around 1050, just slightly above the mathematical midpoint. [VERIFY most recent mean score]

Within your two section scores, there are also subscores and cross-test scores that break down your performance further. You'll see subscores for things like "Command of Evidence," "Words in Context," "Heart of Algebra," and "Passport to Advanced Math." You'll also see cross-test scores for "Analysis in History/Social Studies" and "Analysis in Science," which pull questions from across both sections. These subscores range from 1-15 or 10-40 depending on the category. [VERIFY current subscore ranges for digital SAT]

Here's what matters about subscores: most colleges don't look at them. The vast majority of admissions offices care about your total score and, in some cases, your section scores. A few engineering programs might glance at your Math section specifically. A handful of highly selective schools might note a significant gap between your EBRW and Math scores. But the subscores? Those are diagnostic tools for you, not evaluation tools for them. They're useful for figuring out where to focus your prep and mostly irrelevant for applications.

The College Board's SAT Technical Manual explains the reliability of scores in terms of Standard Error of Measurement, or SEM. For the total score, the SEM is roughly 40 points. [VERIFY current SEM for total score] That means if you scored a 1200 and took the test again under identical conditions — same preparation, same sleep, same breakfast — your score would fall between about 1160 and 1240 roughly two-thirds of the time. This is why a 10-point difference between two scores is pure noise. It's meaningless. A 20-point difference is still within the margin of error. You need to see a gap of about 60 points or more before you can confidently say one score is actually higher than the other.

The Math

Let's walk through the conversion with rough numbers so you can see how this works. Say you take the Math section and answer 45 out of 58 questions correctly. That's your raw score: 45. The equating table for your particular test form might convert that to a scaled Math score of 650. On a different test form — one that happened to be slightly easier — a raw score of 45 might only convert to 630. Same number right, different scaled score. You didn't do anything differently. The test was different.

This is why chasing a specific score number can feel maddening. You can improve your actual ability, answer more questions correctly, and still see your scaled score stay flat — if the form you took was easier than the previous one. The inverse is also true: you might score higher on a harder form with the same raw performance. The equating process is designed to wash all of this out over time, but on any single test day, it can feel arbitrary.

For the essay section — this is relevant if you're looking at older score reports or if anyone brings it up. The College Board removed the optional essay in 2021. Before that, it was scored separately on three dimensions (Reading, Analysis, Writing) by two human graders, each giving 1-4 on each dimension, for a total range of 2-8 per dimension. It was never folded into your composite score. Many colleges never required it even when it existed. If someone tells you that you need to worry about the SAT essay, they're working with outdated information.

One more number worth knowing: score choice. The College Board lets you decide which test date's scores to send to colleges through a policy called Score Choice. Some colleges accept it; others require you to send all scores. When colleges accept Score Choice, they typically practice "superscoring" — taking your highest EBRW from any sitting and your highest Math from any sitting and combining them into a new composite. This means every time you retake the SAT, you're only risking time and energy, not your best score. If a college superscores, a retake can only help you. The College Board's website lists their Score Choice policy, and individual colleges publish their superscoring policies on their admissions pages. [VERIFY that Score Choice is still current policy]

What Most People Get Wrong

The most common misunderstanding is treating your score like a grade on a test you either passed or failed. It's not. It's a position on a statistical distribution that was engineered to produce a bell curve. The scale, the equating, the SEM — it's all built to rank you relative to other test-takers, not to measure some absolute threshold of knowledge. When you understand that, you stop thinking "I got a 1150, I must not know enough" and start thinking "I placed at the 67th percentile of the people who took this test this year, and here's what I'd need to do to move up."

Another mistake is obsessing over small score changes on retakes. If you scored a 1180 the first time and a 1200 the second time, you didn't meaningfully improve. That 20-point change is within the standard error of measurement. You might have just had a better night's sleep. Real improvement — the kind that changes your admissions picture — looks like 60 points or more, and that usually requires targeted preparation, not just showing up again and hoping.

People also get confused about section balance. A 1300 made up of 700 EBRW and 600 Math is not the same as a 1300 made up of 600 EBRW and 700 Math, at least not for every school. If you're applying to a liberal arts college, the first breakdown might actually serve you better. If you're aiming at an engineering program, the second is stronger. Same total, different story. Knowing your section scores and understanding what they mean for your specific target schools is more useful than fixating on the composite.

Finally, don't let anyone tell you the test is "easier now" because there's no wrong-answer penalty or because the essay is gone. The scoring scale adjusted to account for these changes. A 1300 today is not a 1300 from 2005 — the tests are different instruments with different equating tables. Comparing scores across eras is comparing apples to orangutans. Focus on what your score means in the context of the students taking the test alongside you, because that's exactly what colleges are doing.

This article is part of the SAT Real Talk series at SurviveHighSchool.

Related reading: What the SAT Actually Measures (It's Not Intelligence), What Your Score Range Actually Means for College, You Got an 1100. Here's What Happens Next.