How this percentile calculator works
Give it a list of numbers and it answers either of the two questions people bring to a percentile. Ask for the value at a percentile (where does the 90th percentile of these scores fall?) and it computes the answer by all three standard methods, side by side, with the rank arithmetic shown for each. Ask for the percentile of a value (my score was 29, where does that sit?) and it counts your list the standard way: everything below you, plus half of anything that ties you.
Why three methods? Because there is genuinely no single agreed formula for a percentile of a finite list, and the big tools disagree with each other. Excel alone ships two. This page would rather show you all three and name who uses which than quietly pick one and leave you wondering why your textbook says something else.
The formulas
Inclusive: position = (n - 1) × p ÷ 100 + 1, then interpolate
Exclusive: position = (n + 1) × p ÷ 100, then interpolate
Percentile rank of a value: (count below + 0.5 × ties) ÷ n × 100
Here p is the percentile you asked for and n is how many values you have. "Interpolate" means: when the position lands between two sorted values, slide the matching fraction of the way from the lower one to the upper one. Nearest rank never interpolates, which is why its answer is always a value that actually appears in your data.
Worked example: one list, three answers
Take the data set 15, 20, 35, 40, 50 and ask for the 40th percentile.
Nearest rank computes ceiling of (40 ÷ 100 × 5) = 2 and takes the 2nd smallest value: 20.
Inclusive interpolation (Excel's PERCENTILE.INC) computes position (5 - 1) × 0.4 + 1 = 2.6, so it goes 60% of the way from 20 to 35: 29.
Exclusive interpolation (Excel's PERCENTILE.EXC) computes position (5 + 1) × 0.4 = 2.4, so it goes 40% of the way from 20 to 35: 26.
Same five numbers, same question, three answers: 20, 26, or 29, and every one of them is computed correctly by its own rule. Ask the same list for its 50th percentile instead and all three methods land on 35, which is also the median. Agreement like that is common at friendly percentiles on small round data sets, and its absence is usually the whole reason your answer differs from your textbook's.
Why Excel, your textbook, and your stats package disagree
A percentile is easy to define for a smooth curve and genuinely underdetermined for a short list. Where exactly is the 40th percentile of five numbers? Somewhere around the 2nd and 3rd values, certainly, but any rule that picks a precise spot is a convention, not a discovery. Three conventions won. The nearest rank rule is the classic textbook definition, and it always returns an actual data value. Inclusive interpolation treats the smallest value as the 0th percentile and the largest as the 100th; it is Excel's PERCENTILE.INC (and plain PERCENTILE), Google Sheets, and the numpy default. Exclusive interpolation reserves a little room beyond both ends, on the sensible argument that a sample probably does not contain the true extremes; it is Excel's PERCENTILE.EXC, the default in Minitab, and the closest cousin of how TI calculators build their quartiles.
Nobody here is wrong, and the differences fade as the list grows: on a thousand values the three answers typically agree to within the gap between neighboring data points. On a class of 12 quiz scores they can differ visibly, which is exactly when students meet them. So if your answer and your tool's answer differ, check the method before checking your arithmetic. It is nearly always the method.
A percentile is a position, not a score
The 99th percentile on a 60-question test might be 51 correct answers. A percentile says nothing about how much of the material you knew; it says how much of the group sits at or below you. That makes percentiles wonderful for comparing and useless for grading: everybody could do well, or everybody could struggle, and the percentiles would come out the same either way.
It also settles a question that keeps new parents up at night. On a growth chart, a baby at the 10th percentile for weight is not failing at anything. Ten percent of perfectly healthy babies live there, by definition; somebody has to, or the chart would be wrong. What a pediatrician actually watches is not the number but the track: a child who follows their own curve, whichever curve it is, is doing exactly what growth charts hope to see, while a rapid slide across several percentile lines is the pattern that earns a closer look. And since half of everyone is below average on any given measure, being below the 50th line somewhere is not a finding. It is arithmetic.
Percentile, percentile rank, and percentage
Three words, three jobs. A percentage is a fraction of a quantity: 40% of 50 points is 20 points, and our percentage calculator handles that family. A percentile is a value that marks a position in a group: the 40th percentile of the five numbers above is 29 by Excel's count. A percentile rank runs the same idea backwards: given a value, it reports the position, which is this page's second mode. The 50th percentile has its own name, the median, and quartiles are the 25th, 50th and 75th percentiles wearing different badges. Our quartile calculator specializes in those, and our mean, median and mode calculator covers the centers.
When you have a mean and a standard deviation instead of a list
IQ scores, standardized tests and lab reference ranges usually hand you a mean and a standard deviation rather than a raw list. That percentile comes from the area under a normal curve, which is a different computation from counting a list, and our z-score calculator does it properly. This page needs your actual numbers; that one needs only the two summary figures. And if you want the spread of your list itself, the standard deviation calculator and the average calculator take the same paste-in list this page does.