SAT percentiles explained

Your SAT score has three different percentiles attached to it, and only two of them have a published lookup table. Most sites that answer “what percentile is my score” quietly pick one of the two published tables and present it as the answer, which is why the number on a student’s own score report often fails to match anything found online.

Here is the actual structure. Two percentile systems are published in College Board’s research tables. A third, different number is what appears on the score report itself.

The three systems side by side

System	Comparison group	Published table
Nationally Representative Sample	All US 11th and 12th graders, modelled from a research study	Yes
SAT User Group	Students who graduated in the past three school years and took the SAT	Yes
All Tester Percentile	Past three cohorts of 12th grade SAT takers worldwide	No

Nationally Representative Sample

College Board describes this one as “derived from a research study of U.S. students in 11th and 12th grade and are weighted to represent all U.S. students in those grades, regardless of whether they typically take the SAT” (College Board Research).

The key phrase is “regardless of whether they typically take the SAT”. This group includes students who never sit the exam, so it is a wider and generally lower-scoring reference population. That is why a mid-scale score usually earns a higher percentile here than in the User table.

SAT User Group

This one is “based on the actual SAT scores of students that graduated in the past three school years”. Everyone in this group actually took the test, which makes it a self-selected population, and self-selection varies by state and by school policy.

Because it is a rolling three-year window, this table changes when a new graduating class enters it and the oldest class drops out. A score can gain or lose a percentile point without the student doing anything.

All Tester Percentile, the one on your report

The current student score report shows neither of the tables above. It shows the All Tester Percentile, which College Board’s Fall 2026 Understanding Scores document defines as being “based on the actual scores of the past 3 cohorts of students who took the SAT in 12th grade for tests completed anywhere in the world” (Understanding Scores PDF).

Two features of that definition matter. First, it is worldwide, not US only. Second, it counts students who tested in 12th grade, which is a narrower filter than “graduated in the past three school years”.

College Board publishes no lookup table for the All Tester Percentile. No calculator on any site, including ours, can reproduce it exactly. Any tool that claims to match the number on your report is quietly showing you a User Group or Nationally Representative figure with a different label. The SAT percentile calculator here labels which published table each figure comes from, and says plainly that neither is the report number.

Online score insights add two further comparisons, a Country percentile and a State percentile. Those are also unpublished as tables.

The published percentile table

Both published systems, in 50 point steps. Full ten point resolution is in the SAT percentile reference.

Total score	Nationally Representative	SAT User Group
1600	99+	99+
1550	99+	99
1530	99+	99
1500	99	97
1450	99	95
1400	97	93
1360	95	90
1300	91	85
1250	86	80
1200	81	75
1150	74	69
1100	67	62
1050	58	55
1020	52	50
1000	48	47
950	38	39
900	29	32
860	23	26
850	21	24
800	14	17
750	8	11
700	4	6
650	1	3
600	1-	2

Notice how the two columns cross over. Below about 1000 the User percentile runs higher, and above it the Nationally Representative percentile runs higher. That is the self-selection effect visible in a single table: the population that chooses to take the SAT is stronger on average than the population that does not.

How to read a percentile without misreading it

Percentiles are inclusive. College Board defines them as “at or below”, so a 50th percentile score means half the comparison group scored at that score or lower, students at exactly that score included. This is why the mid-scale percentile sometimes appears to jump more than a point per ten point step: a lot of students cluster on the same scores.

The printed values “99+” and “1-” are censored bounds, not computed numbers. College Board prints them that way rather than resolving the extreme tails, and any site showing “99.7th percentile” for a 1580 has invented the precision.

Percentile is a rank, not a score, and the exchange rate between them is not constant. In the User Group table, 1020 is the 50th percentile and 1100 is the 62nd, so 80 points buys 12 percentile points in the middle of the scale. Near the top, 1500 is the 97th and 1530 is the 99th, where 30 points buys 2. An improvement target set in percentile terms means a very different number of points depending on where you start. Our methodology page documents which table each tool draws from.

A percentile is not a percentage, and not a curve

Three things get confused with each other, and they are unrelated measurements.

A percentile is a rank against other students. A percentage correct is your own answer count divided by the question count. They can differ wildly, because 90 scored questions on the digital SAT are weighted by difficulty under Item Response Theory rather than counted flat, as set out in how the digital SAT is scored.

Neither is a curve. The SAT is not scored relative to the people who sat the same date, and a strong or weak test day nationally does not move your score. Percentiles are computed after the fact from a fixed reference group, so they describe where your score lands among past cohorts rather than shaping the score itself.

Section percentiles are separate tables

Reading and Writing and Math each have their own percentile distributions, and they are not identical. A 600 sits at the 72nd percentile in Reading and Writing and the 75th in Math among User Group students.

Section score	RW (Nat. Rep.)	RW (User)	Math (Nat. Rep.)	Math (User)
800	99+	99+	99+	99+
750	99	98	98	95
700	97	92	95	91
650	90	84	90	84
600	79	72	81	75
530	58	53	61	59
500	48	43	47	50
480	41	37	40	45
400	16	16	15	21

Section percentiles do not add up to the total percentile. The total table is built from the distribution of total scores, not from combining section ranks.

What a percentile is useful for

A percentile answers one narrow question: how common is this score within one defined group. It does not carry admissions information, because colleges compare applicants against their own pools rather than the national one. For that comparison, use each college’s published middle 50 percent range and the score needed calculator.

The PSAT/NMSQT has its own percentile tables on a narrower 320-1520 scale, and its Selection Index is a separate calculation entirely. Those are handled by the PSAT score calculator and the National Merit Selection Index calculator.

For where each number on this page came from, see data sources. For the cohort figures behind the User table, see the average SAT score.