A percentile calculator finds the value below which a given share of your data falls, and the percentile rank of a value in the data. Paste your numbers, choose a method, and enter a percentile such as 90 to get the value, along with the quartiles and deciles. Or enter a value to see its percentile rank. It offers the nearest-rank method and two forms of linear interpolation, since textbooks and software disagree. It runs in your browser.
How to use the percentile calculator
- Choose value at a percentile or percentile rank.
- Paste the data.
- Choose the method and enter the percentile, or the value.
- Press Calculate.
The nearest-rank method
The Wikipedia article on percentiles defines the P-th percentile by the nearest-rank method as the smallest value in the list such that at least P percent of the data is less than or equal to it. The rank is n = ⌈(P ÷ 100) × N⌉, and you take the value at that place in the sorted list. The 100th percentile is the largest value.[1] For the list 15, 20, 35, 40 and 50, the 30th percentile has rank ⌈1.5⌉ = 2, which is 20, the 40th has rank 2, which is also 20, and the 50th has rank ⌈2.5⌉ = 3, which is 35. This method always returns a value that is in the data.
Linear interpolation
The page also describes interpolation, in which the value is found at a fractional rank, between two neighbouring values: x = (N + 1 − 2C) × p + C, where p is the percentile as a fraction. With C = 1, the rank is p(N − 1) + 1, which is the inclusive method used by NumPy and by Excel's PERCENTILE.INC. With C = 0, the rank is p(N + 1), the exclusive method that NIST recommends and that Excel's PERCENTILE.EXC uses, and it is not defined for percentiles close to 0 or 100.[1] Between two values, the answer is the lower value plus the fractional part of the rank times the gap. For the list 15, 20, 35, 40, 50 and the 30th percentile, the inclusive rank is 2.2, giving 20 + 0.2 × 15 = 23, and the exclusive rank is 1.8, giving 15 + 0.8 × 5 = 19.
Matching the page's example
The page has a table for a list of 10 scores. For the 10th, 25th, 50th, 75th and 90th percentiles, the exclusive method gives 1.1, 2.75, 3.5, 5 and 6.8, and the inclusive method gives 1.9, 3, 3.5, 4.75 and 5.2.[1] The page's own list is not shown in the part that we read, but the list 1, 2, 3, 3, 3, 4, 4, 5, 5 and 7 reproduces every one of these values, and the calculator gives them for it. The page says that there is no standard definition of a percentile, which is why you must state the method.
Quartiles and deciles
The page maps the 25th percentile to the first quartile, the 50th to the median, or second quartile, and the 75th to the third quartile.[1] The calculator shows them for your chosen method. The deciles, which divide the data into tenths, are the 10th, 20th, up to the 90th percentiles. The page links to a separate article on deciles and gives no definition, so the calculator takes the kth decile as the 10k-th percentile, which is the usual one.
Percentile rank
The percentile rank of a score is the share of the data that falls below it. The page gives PR = (CF − 0.5F) ÷ N × 100, where CF is the number of scores at or below the score, F is the number equal to it and N the number of all scores. An equal form is (CF′ + 0.5F) ÷ N × 100, where CF′ counts the scores below it.[2] Its example uses 10 scores. A score of 4 has five scores below it and two equal to it, so its rank is (5 + 1) ÷ 10, or 60 percent. A score of 7 has nine below and one equal, so 95 percent, and a score of 6 has nine below and none equal, so 90 percent.[2] The calculator shows the rank and the share strictly below and at or below the value.
Choosing a method
Use nearest rank when you want a value that exists in the data, for example for a service level in a call centre. Use the inclusive interpolation to match Excel's older function and most statistics libraries. Use the exclusive form for small samples when you want percentiles that stay inside the range of the data, as some statistics books do. For large data sets, all three methods give nearly the same result.
Using percentiles
Percentiles describe where a result stands. A child at the 90th percentile of height is taller than 90 percent of children of the same age. Exam boards use percentiles to compare candidates, and businesses use the 95th or 99th percentile of response times to describe worst-case service. The median, the 50th percentile, is a robust summary that is not much changed by extreme values.
Common mistakes
- Forgetting to sort. The calculator sorts for you.
- Mixing up percentile and percentage. A score of 80 percent need not be at the 80th percentile.
- Using a different method from your course. Check which one is required.
- Using the exclusive method near the ends. It is not defined there.
What next
Use this often? Add Percentile to your Home screen.
Frequently asked questions
What is a percentile calculator?
A tool that finds the value below which a given share of your data falls, and the percentile rank of a value.
What is the nearest-rank method?
The P-th percentile is the value at rank ⌈P ÷ 100 × N⌉ in the sorted list. The 100th percentile is the largest value.
What is the difference between PERCENTILE.INC and PERCENTILE.EXC?
INC uses the rank p(N − 1) + 1 and EXC uses p(N + 1). EXC is not defined for percentiles close to 0 or 100.
How is percentile rank calculated?
(the number below the score + half of the number equal to it) ÷ N × 100.
What are deciles?
The values that divide the data into ten equal parts: the 10th, 20th and so on to the 90th percentile.
Was this tool helpful?
What went wrong?
We only receive the reason you pick, never what you typed into the tool.
Thanks for your feedback.