Percentile Rank Calculator

A percentile is a cut point in the dataset: the 90th percentile is the value below which 90 percent of the data falls. A percentile rank goes the other direction: you start with a value and ask what percent of the dataset sits at or below it. A test score at the 73rd percentile rank means roughly 73 percent of scores were lower or equal. The percentile slices the data; the percentile rank locates a single value within the data.

Exclusive uses belowCount / N times 100, counting only values that are strictly less than yours. Inclusive uses (belowCount + equalCount) / N times 100, counting values that are less than or equal. Midpoint splits the difference: (belowCount + 0.5 times equalCount) / N times 100. This calculator also offers an Excel inclusive method, which interpolates the sorted input. The methods can differ even without ties, and ties can widen the gap.

NIST and Excel do not share one midpoint default. NIST documents several percentile definitions, and Excel's PERCENTRANK.INC and PERCENTRANK.EXC use their own interpolation rules. Midpoint is an explicit option in this calculator that treats tied values symmetrically. Match the method required by your data source, rubric, or software.

The methods can differ even when every value is unique. Ties add another choice: exclusive ignores equal values, inclusive counts every equal value, and midpoint counts half. The larger the cluster of identical values at your number, the wider the gap can become. If a source specifies a method, use that method rather than choosing one from the result.

Standardized test reports (SAT, GRE, GMAT) show your score and the percentile rank of that score across all test takers. Pediatric growth charts use height and weight percentile ranks to compare a child against the reference population for their age. Salary surveys publish the 25th, 50th, and 75th percentile rank for each role. Sports rankings, college admissions cutoffs, fitness benchmarks, and clinical lab reference ranges all rely on the same idea: a single number is most useful when you know how it stacks up against everyone else.