A percentile rank is the percentage of scores in a comparison group that fall at or below a given score — it is not the percentage of questions answered correctly. In a 15-student class where 7 students scored 73 or below out of 15 total scores, 73 sits at the 46.7th percentile (7 ÷ 15 × 100), even though 73 might be a 73% raw score on the test itself; the two numbers answer different questions and can differ by a wide margin.
What is the worked percentile rank for a score of 73?
A class of 15 students has sorted raw scores 58, 62, 65, 68, 70, 71, 73, 75, 78, 80, 82, 85, 88, 90, 95. To find the percentile rank of the score 73, count how many of the 15 scores are less than or equal to 73: that is 58, 62, 65, 68, 70, 71, and 73 itself — 7 scores. Divide 7 by the total group size, 15, and multiply by 100: 7 ÷ 15 × 100 = 46.67, so a raw score of 73 sits at the 46.7th percentile in this group — just under the middle, even though 73 looks like a solid raw score on a 100-point scale. A higher score in the same group, 85, has 12 of the 15 scores at or below it, giving 12 ÷ 15 × 100 = 80.0 — the 80th percentile.
How do I calculate a percentile rank myself?
Sort the comparison group’s scores from lowest to highest, count how many scores are less than or equal to the target score (including the target score itself), divide that count by the total number of scores in the group, and multiply by 100. The percentile rank changes if the comparison group changes, even when the raw score itself does not — a score in a stronger group falls to a lower percentile, and the same score in a weaker group rises to a higher one. The percentage calculator checks the final division step; the standard deviation calculator shows how spread out a group’s scores are around their average, a related but separate measure from percentile rank.
How do I check a percentile-rank result?
A useful check is that the percentile rank of the lowest score in a group can never be 0 and the percentile rank of the highest score is always 100, since each score counts itself in the “at or below” tally — in the 15-score class, the lowest score, 58, is at or below itself, giving 1 ÷ 15 × 100 = 6.7, and the highest, 95, gives 15 ÷ 15 × 100 = 100. A second check: percentile rank is monotonic with raw score inside one comparison group — a higher raw score can never produce a lower percentile rank than a lower raw score from the same group, so a computed table should always read as a non-decreasing sequence when scores are sorted.
What mistakes produce a wrong percentile rank?
The most common mistake is treating a percentile rank as a percentage-correct raw score — “85th percentile” does not mean 85% of questions answered correctly; it means the test-taker scored at or above 85% of the comparison group, whatever the actual percentage-correct figure was. A second mistake is comparing percentile ranks computed from two different comparison groups as though they were on one shared scale — a student in the 60th percentile of their own class and a student in the 60th percentile of a national norm group used entirely different denominators, even though both figures say “60th percentile.” A third is dividing by the wrong group size, such as using the number of questions on the test instead of the number of test-takers in the comparison group.
What does a percentile rank not establish?
This method computes a percentile rank from a stated, finite comparison group of raw scores. It does not itself establish that the comparison group is representative of a wider population, correct for how a test was equated across different versions or years, or account for measurement error in an individual score — those require additional information from the test publisher’s scoring documentation, not from the percentile arithmetic alone.
Source for the external fact
The College Board explains that a percentile compares a test-taker’s score against the scores of a specific reference group of other test-takers, and that a percentile of, for example, 75 means the test-taker scored as well as or better than about 75% of that reference group, in its SAT score-reporting documentation.
A reference table across the 15-score class
The same at-or-below count, divided by 15, produces every row.
| Raw score | Scores at or below | Percentile rank |
|---|---|---|
| 58 | 1 | 6.7 |
| 68 | 4 | 26.7 |
| 73 | 7 | 46.7 |
| 80 | 10 | 66.7 |
| 85 | 12 | 80.0 |
| 95 | 15 | 100.0 |
A second worked example with a larger, evenly spaced group of 20 scores
A group of 20 scores runs consecutively from 40 to 59 with no gaps. For a target score of 52: scores at or below 52 are 40 through 52 inclusive, which is 13 scores. Percentile rank: 13 ÷ 20 × 100 = 65.0, the 65th percentile. Because this group is evenly spaced with no ties, its percentile ranks increase by a constant 5 percentage points for every one-point increase in raw score (1 ÷ 20 × 100 = 5) — a pattern that does not hold for the first, unevenly spaced 15-score class above, where the gaps between consecutive raw scores vary.
The tied-score edge case, where several test-takers share one raw score
A small group of 7 scores is 10, 12, 12, 12, 15, 18, 20 — three test-takers tied at 12. For the tied score of 12, the “at or below” count includes all three tied scores plus the one score below them (10): 4 scores at or below 12 out of 7 total, giving 4 ÷ 7 × 100 = 57.1, the 57.1th percentile — the same percentile rank assigned to all three test-takers who scored 12, even though a naive count that only located “the” position of 12 in the sorted list might undercount by including just one of the three ties. Consistently counting every tied instance in the “at or below” tally, not just one, keeps the method correct when scores repeat.
Why percentile rank and raw score answer different practical questions
A raw score of 73 out of 100 answers “how many of the possible points did this test-taker earn,” independent of anyone else’s performance — it stays 73 whether the test was easy or hard, and whether the rest of the group did well or poorly. A percentile rank of 46.7 answers “how did this test-taker do relative to this specific group,” and would change if the same test-taker retook an identically scored test alongside a stronger or weaker group of peers, even though their raw score stayed at 73. A school report, a job screening test, or a standardized exam typically publishes both figures side by side precisely because neither one substitutes for the other.
Class percentile vs a national norm group: the identical raw score, two different percentile ranks
The 15-student class average used throughout this guide is one comparison group, but a large standardized test also reports a percentile against a much larger national norm group — and the same raw score can land at two noticeably different percentile ranks depending on which group it is compared against. A student scoring 73 who ranks at the 46.7th percentile within their own 15-student class might rank at, say, the 62nd percentile against a national norm group if that class happens to be somewhat stronger on average than the national sample — or lower than 46.7 if the class is weaker than average. Neither percentile figure is “more correct” than the other; each answers a legitimate but different question, and a school report that quotes only one of the two without naming its comparison group leaves out the information needed to interpret the number correctly.
Related calculators
Check a percentage or ratio result with the percentage calculator, measure how spread out a group’s scores are with the standard deviation calculator, and see how population vs sample counting changes that spread figure in the population vs sample guide.