Linear P90
For 10, 20, 30, 40, 50, linear inclusive interpolation places P90 at position 3.6 (zero-based), giving 46.
Percentile Calculator — Rank, Quartiles & Percentile Values
Dataset and questions
Results
Results will appear here.
Enter a dataset and select Calculate.
Advertisement
Percentile formulas and methods
Sort the observations as x(1) ≤ … ≤ x(n). A sample percentile estimates a position in this ordered list, while percentile rank asks where a particular target sits relative to the observations.
| Quantity or method | Calculation | Interpretation |
|---|---|---|
| Linear inclusive (R-7) | h = (n − 1)p; interpolate between zero-based positions floor(h) and ceil(h) | Includes the sample minimum at p = 0 and maximum at p = 1. This matches Excel PERCENTILE.INC and NumPy's default linear method. |
| Nearest rank | r = ceil(np), with r = 1 when p = 0 | Returns an observed value rather than an interpolated value. |
| Midrank percentile rank | 100 × (B + 0.5E) / n | B is the count below the target and E is the count equal to it. |
| Strict percentile rank | 100 × B / n | Counts only observations strictly below the target. |
| Inclusive percentile rank | 100 × (B + E) / n | Counts observations below or equal to the target. |
| Quartiles and IQR | Q1 = P25, Q2 = P50, Q3 = P75; IQR = Q3 − Q1 | Uses the selected percentile-value method consistently. |
Method note: Percentile conventions vary. The NIST percentile reference describes ordered data and common percentile estimation, while NumPy documents the available quantile methods. Always report the method when results must be reproducible.
How to use the percentile calculator
- Paste the dataset. Commas, spaces, semicolons, tabs, and line breaks are accepted.
- Enter a value to rank. It does not need to occur in the dataset.
- Enter a percentile to find from 0 to 100, such as 90 for P90.
- Choose the percentile-value and percentile-rank methods. Keep the method label with any result you share.
- Select Calculate. Copy or download the text summary if useful.
Percentile and quartile examples
Nearest-rank P90
For the same five values, ceil(5 × 0.90) = 5, so nearest-rank P90 is the fifth observed value: 50.
Ties in percentile rank
In 10, 20, 20, 40, the midrank of 20 is 100 × (1 + 0.5 × 2) / 4 = 50%.
How to interpret percentile rank, percentiles, and quartiles
A percentile value is expressed in the same units as the data. If P90 is 46 seconds, then 46 seconds is the estimated 90th-percentile threshold under the selected method. A percentile rank is a percentage describing a target's relative position. These are inverse questions and are not always exact inverses in a finite sample, especially with ties or a discontinuous method.
Q1, Q2, and Q3 divide the ordered scale around the 25th, 50th, and 75th percentiles. The IQR measures the middle half's spread and is less sensitive to extreme values than the full range. Percentiles summarize the data entered; they do not by themselves establish probability, causation, or population representativeness.
Frequently asked questions
What is a percentile?
A percentile is a data value associated with a stated percentage of an ordered dataset. For example, the 75th percentile is a value near or above which roughly one quarter of observations lie, subject to the chosen method.
What is percentile rank?
Percentile rank describes a target's relative standing as a percentage of observations below it, at or below it, or below it with half of ties counted, depending on the selected convention.
How does the calculator handle ties?
The default midrank convention counts every lower value plus half of the values equal to the target. You can instead count only values strictly below or all values at or below the target.
How are Q1, Q2, and Q3 calculated?
Q1, Q2, and Q3 are the 25th, 50th, and 75th percentiles under the selected percentile-value method. IQR equals Q3 minus Q1.
Why can percentile answers differ between calculators?
Sample percentiles do not have one universal estimation rule. Software can use different rank positions, interpolation, endpoints, and tie conventions. This calculator labels the method used.
Can I paste spreadsheet data?
Yes. Paste a row, column, or block separated by commas, spaces, semicolons, tabs, or line breaks. Do not use commas as thousands separators because commas are treated as delimiters.
What happens with one data value?
Every percentile value is that sole observation. The target's percentile rank still depends on whether it lies below, equals, or exceeds that observation and on the selected rank method.
Are my data uploaded or stored?
No. The dataset and calculations stay in your browser; this tool does not upload or store the values you enter.