Knowing how to calculate percentiles and percentile rank lets you turn a raw number into a position: where a score, salary, height or response time sits compared with everyone else in the group. A score of 78 means little on its own. Knowing it beats 52% of the group tells you much more. With the formulas below you can work it out by hand, and then check the result with a calculator on Calculatorr.
What is a percentile and what is percentile rank?
A percentile is a value below which a given percentage of observations fall. The 90th percentile is the value that 90% of the data points are at or below. Percentile rank works the other way around. You start with a value and ask what percentage of the data sits below it.
In short:
- Percentile: you know the percentage and want the value (for example, the 75th percentile of a set of salaries).
- Percentile rank: you know the value and want the percentage (for example, what percentile a score of 78 is).
Both depend on the data set you compare against. The same score can have a different percentile rank in different groups.
Percentile rank formula
The most common formula for percentile rank is:
PR = ((L + 0.5 × S) ÷ N) × 100
- L: number of values lower than the value you are ranking.
- S: number of values equal to it (including itself).
- N: total number of values.
The 0.5 × S term splits tied values evenly between "below" and "above". If there are no ties, S is 1 and the formula still works. Some textbooks use a simpler version, PR = (L ÷ N) × 100, which counts only the values strictly below. Both are valid. What matters is that you state which one you use and apply it consistently.
How to calculate percentile rank step by step
Take the scores of 20 students on an exam, already sorted from lowest to highest:
55, 58, 62, 64, 67, 70, 71, 73, 75, 76, 78, 80, 82, 84, 85, 87, 89, 91, 94, 97
What is the percentile rank of a student who scored 78?
Step 1: Sort the data
Sorting makes counting easy and prevents mistakes. The list above is already in ascending order.
Step 2: Count the values below the score
Ten scores are lower than 78 (from 55 up to 76), so L = 10.
Step 3: Count the values equal to the score
Only one student scored 78, so S = 1.
Step 4: Apply the formula
PR = ((10 + 0.5 × 1) ÷ 20) × 100 = (10.5 ÷ 20) × 100 = 52.5
The student sits at the 52.5th percentile. With the simpler formula, the result would be (10 ÷ 20) × 100 = 50.
Example with tied values
Imagine five scores: 70, 75, 75, 80, 90. To find the percentile rank of 75, count L = 1 (only 70 is lower) and S = 2 (two scores equal 75).
PR = ((1 + 0.5 × 2) ÷ 5) × 100 = (2 ÷ 5) × 100 = 40
Without the tie correction you would get 20%, which understates the position of both students who scored 75.
How to calculate a percentile from a data set
To find the value at a given percentile, you first locate its position (rank) in the sorted list and then read the value. There are three common ways to do it, and they give slightly different answers.
Nearest-rank method
n = ⌈(P ÷ 100) × N⌉
Round up to the next whole number and take the value at that position. For the 90th percentile of our 20 scores: 0.90 × 20 = 18, so the answer is the 18th value, which is 91.
Inclusive method with interpolation
Rank = (P ÷ 100) × (N − 1) + 1
This is the logic behind PERCENTILE.INC in spreadsheets. For the 90th percentile: 0.90 × 19 + 1 = 18.1. The 18th value is 91 and the 19th is 94, so the result is 91 + 0.1 × (94 − 91) = 91.3.
Exclusive method with interpolation
Rank = (P ÷ 100) × (N + 1)
This matches PERCENTILE.EXC. For the 90th percentile: 0.90 × 21 = 18.9, so the result is 91 + 0.9 × (94 − 91) = 93.7.
| Method | Rank position | 90th percentile |
|---|---|---|
| Nearest rank | 18 | 91 |
| Inclusive (N − 1) | 18.1 | 91.3 |
| Exclusive (N + 1) | 18.9 | 93.7 |
With large data sets the three methods converge. With small ones the differences can matter, so always note which one you used.
More examples with the same data
- 25th percentile (inclusive): rank = 0.25 × 19 + 1 = 5.75. The 5th value is 67 and the 6th is 70, so the result is 67 + 0.75 × 3 = 69.25.
- 50th percentile (inclusive): rank = 0.50 × 19 + 1 = 10.5. The 10th and 11th values are 76 and 78, so the result is 77. This is the median.
- 25th percentile (nearest rank): 0.25 × 20 = 5, so the answer is the 5th value, 67.
Percentile vs percentage vs quartile: what is the difference?
These terms are often confused, and mixing them up leads to wrong conclusions.
| Term | What it measures | Example |
|---|---|---|
| Percentage | A part of a total | Answering 78 of 100 questions correctly is 78% |
| Percentile | Position within a group | Beating 78% of test takers is the 78th percentile |
| Quartile | Position in four equal parts | Q1 is the 25th percentile, Q3 is the 75th |
Someone can score 40% on a very hard test and still be at the 90th percentile if almost everyone else scored lower. For a refresher on the first concept, see how to calculate percentages on Calculatorr.
How to interpret percentile results
A percentile tells you rank, not size. A few guidelines help you read the number correctly:
- Around the 50th percentile: typical, at the middle of the group.
- Above the 75th percentile: in the top quarter.
- Below the 25th percentile: in the bottom quarter.
- Above the 90th or 95th: clearly high compared with the group, and the cutoff many programs use to flag outstanding results.
In our example, the student at the 52.5th percentile did slightly better than half the class. That is solid but average, even though 78 may look like a good grade. Percentile gaps are also uneven. Going from the 50th to the 60th percentile may need only a couple of points, while going from the 95th to the 99th may need many more, because data is dense around the middle and sparse at the extremes.
Where percentiles are used in real life
- Exams and admissions: standardized tests report percentile rank so you can compare your score against all test takers.
- Child growth charts: a child at the 40th percentile for height is taller than 40% of children of the same age and sex.
- Salary benchmarking: companies compare pay against the 25th, 50th and 75th percentiles of the market.
- Performance monitoring: engineers track the 95th percentile of response time. If 200 requests are sorted by time, the nearest-rank 95th percentile is the 190th value, since 0.95 × 200 = 190. It shows what slow users experience better than the average does.
How to use a percentile calculator online
When your data set has hundreds of values, doing this by hand is slow. A percentile calculator speeds it up. The workflow is simple:
- Paste or type your values, separated by commas or line breaks.
- Choose what you want: a percentile value (enter the percentage) or a percentile rank (enter the value).
- Select the method if the tool offers one, such as nearest rank, inclusive or exclusive.
- Read the result and compare it with your manual estimate.
You can find this and related tools, such as averages, ratios and percentage calculators, on Calculatorr. Running a few numbers by hand first is a good way to confirm the tool uses the method you expect.
Common mistakes when calculating percentiles
- Not sorting the data. Ranks only make sense in ascending order.
- Confusing percentile with percentage. A 90th percentile does not mean 90% correct.
- Ignoring ties. Repeated values change L and S, so count them carefully.
- Mixing methods. Comparing a nearest-rank result with an interpolated one produces confusing differences.
- Using tiny samples. With fewer than about 20 values, a percentile can jump a lot when one value changes.
- Misreading the extremes. The lowest value has a rank above 0% with the midpoint formula, and the highest is below 100%. Percentile rank by this formula rarely reaches exactly 0 or 100.
- Comparing against the wrong group. A percentile only means something relative to the reference group, such as the same age, grade or region.
Quick checklist to calculate a percentile correctly
- Sort the data from lowest to highest.
- Decide whether you need a percentile (value) or a percentile rank (percentage).
- Pick one method and use it consistently.
- Calculate the rank position or count L, S and N.
- Interpolate when the position falls between two values.
- Interpret the result against the group it came from, and verify it with a calculator.
