Skip to content
Statculator
Descriptive Statistics

Percentile Calculator and Percentile Rank Calculator

Free percentile rank calculator: get the percentile rank of a score, or the value at any percentile (P75, P90, P95) of a data set, with steps and 3 methods.

Free, no sign-up Runs in your browserBy Hira Noor, BS MathematicsUpdated

Try
Find the
20 values

Separate values with commas, spaces or new lines. They do not need to be sorted.

The score or measurement to rank. It does not have to be one of the data values.

Percentile methods differ for small datasets; the results table shows all three. Percentile rank counts ties as half.

Results

Percentile rank of 77
57.5
(B + 0.5E) / n × 100, with B = 11, E = 1, n = 20
% below
55%
11 of 20 values are less than 77
% at or below
60%
12 of 20 values are 77 or less
Count, n
20
Values below x, B
11
Values equal to x, E
1
Values above x
8
% above
40%
Where x sits in the sorted data
rank 12
Minimum and maximum
55, 96

Interpretation: A value of 77 is higher than 11 of the 20 values (55%) and tied with 1, which puts it at a percentile rank of 57.5. 8 values are higher (40%), so it sits in the upper half of this dataset.

Percentile rank of 77 under different definitions

DefinitionFormulaResult
Percentile rank: ties count half (headline result)(B + 0.5E) / n57.5%
Strictly below (“% below”)B / n55%
At or below (cumulative %)(B + E) / n60%
Excel PERCENTRANK.INCB / (n − 1)57.89%
Excel PERCENTRANK.EXC(B + 1) / (n + 1)57.14%

B = values below x, E = values equal to x. Excel shows PERCENTRANK results as a proportion with three decimal places unless you set its significance argument.

Sorted data

RankValueRank / n
1555%
25910%
36215%
46420%
56725%
66930%
77035%
87140%
97345%
107450%
117655%
127760%
137965%
148170%
158375%
168580%
178885%
189090%
199395%
2096100%

Highlighted: where 77 appears in the sorted list.

Step-by-step solution

  1. 1

    Sort the data from smallest to largest

    Sorted (n = 20): 55, 59, 62, 64, 67, 69, 70, 71, 73, 74, 76, 77, 79, 81, 83, 85, 88, 90, 93, 96
  2. 2

    Count the values below x and the values equal to x

    B = number of values < 77 = 11
    E = number of values = 77 = 1
  3. 3

    Apply the percentile rank formula

    Half of the values tied with x are counted as below it, so a value is never ranked entirely above or entirely below its own ties.

    PR = (B + 0.5 × E) / n × 100
    PR = (11 + 0.5 × 1) / 20 × 100 = 11.5 / 20 × 100 = 57.5
  4. 4

    Find the two simpler percentages

    % below = B / n × 100 = 11 / 20 × 100 = 55%
    % at or below = (B + E) / n × 100 = 12 / 20 × 100 = 60%

What is a Percentile Calculator?

A percentile rank calculator tells you what share of a dataset falls below a value: PR = (B + 0.5E) / n × 100, where B values are below x and E equal it. A score of 77 in a class of 20, with 11 lower scores, has a percentile rank of (11 + 0.5) / 20 × 100 = 57.5. Run in reverse, the percentile calculator finds the value at a given percentile: the 90th percentile of the same 20 scores is 90.3 by the Excel and NumPy method.

What the Percentile Calculator does

Choose Percentile rank of a value to use it as a percentile rank calculator: where does a score of 77 stand in this class? You get the percentile rank with ties counted as half, plus the plain “% below” and “% at or below” figures that people often mean when they say percentile, and the matching Excel PERCENTRANK.INC and PERCENTRANK.EXC results. If you only need Q1, the median and Q3, the Quartile Calculator compares the quartile conventions in more detail. If you have a mean and standard deviation instead of raw data, use the Z-Score to Percentile Calculator, or the Percentile to Z-Score Calculator to turn a percentile rank into a z-score.

Choose Value at a percentile to turn a percentage into a data value, the kth percentile of a data set for any k: the 90th percentile of exam scores, the P95 of response times, the 10th percentile of birth weights. The calculator sorts the data, shows the position it reads from, and reports the answer under three definitions side by side (nearest rank, inclusive interpolation and exclusive interpolation), together with P10, P25, P50, P75, P90, P95 and P99 for the method you picked.

How to use the Percentile Calculator

  1. Under Find the, keep Percentile rank of a value (the default) or choose Value at a percentile.
  2. Type or paste your numbers into Data values, in any order. A column copied from Excel or Google Sheets works.
  3. For a percentile value, enter Percentile, P (0 to 100) and choose a Percentile method. Use the inclusive option to match Excel’s PERCENTILE.INC, NumPy and R; use nearest rank if the answer must be an actual data value.
  4. For a percentile rank, enter the Value, x you want to place.
  5. Read the result, then check the comparison table and the highlighted rows of the sorted data to see exactly where it came from.

How to find the 75th or 80th percentile of a data set

These steps use the 20 exam scores from the worked example below, already sorted: 55, 59, 62, 64, 67, 69, 70, 71, 73, 74, 76, 77, 79, 81, 83, 85, 88, 90, 93, 96. The inclusive method is the one Excel’s PERCENTILE.INC, Google Sheets, NumPy and R use by default.

  1. Sort the data from smallest to largest and count it: n = 20.
  2. Write the percentile as a proportion: p = 0.75 for the 75th percentile, 0.80 for the 80th.
  3. Find the position h = (n − 1)p + 1. For P75, h = 19 × 0.75 + 1 = 15.25. For P80, h = 19 × 0.8 + 1 = 16.2.
  4. Read or interpolate. P75 sits a quarter of the way from the 15th value (83) to the 16th (85): 83 + 0.25 × 2 = 83.5. P80 sits 0.2 of the way from the 16th value (85) to the 17th (88): 85 + 0.2 × 3 = 85.6.
  5. State the method. By nearest rank, ⌈0.75 × 20⌉ = 15 and ⌈0.8 × 20⌉ = 16, so P75 = 83 and P80 = 85, both actual scores.
Percentiles of the 20 exam scores under each method
PercentileNearest rank, ⌈pn⌉Inclusive, (n − 1)p + 1Exclusive, (n + 1)p
75th (Q3)8383.584.5
80th8585.687.4
90th9090.392.7
95th9393.1595.85

How to calculate percentile rank, and a percentile from a rank

The percentile rank formula counts the values below x, adds half of any values tied with x, and divides by n: PR = (B + 0.5E) / n × 100. For a score of 88 among the 20 exam scores, 16 scores are lower and one equals it, so PR = (16 + 0.5) / 20 × 100 = 82.5.

If you only know a rank rather than the data, the same formula still works. A value ranked r-th from the bottom of n values, with no ties, has r − 1 values below it, so PR = (r − 0.5) / n × 100. Ranked 5th from the top in a class of 40 means 35 students scored lower: PR = (35 + 0.5) / 40 × 100 = 88.75. Going the other way, the value at percentile rank PR sits at rank r = PR × n / 100 + 0.5, rounded to a whole rank.

Some courses use simpler versions: “% below” is B / n × 100 and “% at or below” is (B + E) / n × 100. For 88 these give 80% and 85%; the tie-splitting formula sits between them. The calculator shows all three, so you can match whichever your textbook uses.

Percentile formula and percentile rank formula

Sort the data so that x₁ ≤ x₂ ≤ … ≤ xₙ and write the percentile as a proportion p = P / 100. The three percentile methods differ only in the position they read.

Nearest rank
r = ⌈p × n⌉, P = xᵣ
Inclusive interpolation (PERCENTILE.INC)
h = (n − 1)p + 1
Exclusive interpolation (PERCENTILE.EXC)
h = (n + 1)p
Interpolating at a fractional position h
P = x⌊h⌋ + (h − ⌊h⌋)(x⌊h⌋₊₁ − x⌊h⌋)
Percentile rank of a value x
PR = (B + 0.5E) / n × 100
Formula symbols
SymbolMeaning
xₖk-th smallest value
nnumber of values
ppercentile as a proportion (0.9 for the 90th)
⌈ ⌉, ⌊ ⌋round up, round down to a whole number
B, Enumber of values below x, number equal to x

The inclusive method reaches the minimum at P = 0 and the maximum at P = 100. The exclusive method only covers percentiles from 1 / (n + 1) to n / (n + 1); outside that range this calculator returns the minimum or maximum and says so, while Excel returns an error.

For nearest rank, P = 0 gives rank 0, which does not exist, so the minimum is used.

Worked example

Worked example: the 90th percentile of 20 exam scores

Problem. Twenty students score 62, 71, 55, 88, 93, 67, 74, 81, 59, 77, 85, 69, 90, 73, 64, 79, 96, 70, 83, 76. Find the 90th percentile, and the percentile rank of a score of 77.

  1. 01Sort: 55, 59, 62, 64, 67, 69, 70, 71, 73, 74, 76, 77, 79, 81, 83, 85, 88, 90, 93, 96 (n = 20).
  2. 02Inclusive interpolation: h = 19 × 0.9 + 1 = 18.1, so P90 = x₁₈ + 0.1 × (x₁₉ − x₁₈) = 90 + 0.1 × (93 − 90) = 90.3.
  3. 03Nearest rank: r = ⌈0.9 × 20⌉ = 18, so P90 = x₁₈ = 90.
  4. 04Exclusive interpolation: h = 21 × 0.9 = 18.9, so P90 = 90 + 0.9 × 3 = 92.7.
  5. 05Percentile rank of 77: B = 11 scores are lower and E = 1 equals it, so PR = (11 + 0.5) / 20 × 100 = 57.5.

Answer. The 90th percentile is 90.3 by the Excel/NumPy default, 90 by nearest rank and 92.7 by the exclusive method. A score of 77 has a percentile rank of 57.5: 55% of the class scored lower and 60% scored 77 or lower.

Percentile, percentile rank and percentage are three different things

A percentage describes one result on its own scale: 77 marks out of 100 is 77%. A percentile is a value in the units of the data, chosen so that a stated share of the data lies at or below it. A percentile rank is the reverse lookup: the share of the data that a particular value beats.

So a student can score 77% on a test and be at the 57.5th percentile rank of the class, while the 90th percentile of the same class is a score of 90.3. Mixing these up is the most common mistake in reading test reports and growth charts.

TermQuestion it answersUnitsExample (the 20 scores above)
PercentageHow much of the maximum was achieved?%77 out of 100 = 77%
Percentile (value)Which value has P% of the data at or below it?Same as the dataP90 = 90.3 marks
Percentile rankWhat share of the data is below this value?0 to 100A score of 77 → 57.5

Percentile methods: nearest rank, inclusive and exclusive interpolation

There is no single agreed definition of a sample percentile, and for small datasets the choices give visibly different answers. Hyndman and Fan (1996) list nine in common use; this calculator covers the three you are most likely to need. All of them converge as n grows. Match the method to the tool you need to agree with:

ToolMethod on this page
Excel PERCENTILE and PERCENTILE.INC, Google Sheets PERCENTILEInclusive interpolation, (n − 1)p + 1
NumPy percentile and pandas quantile (default “linear”), R quantile (default type 7)Inclusive interpolation, (n − 1)p + 1
Excel PERCENTILE.EXC, Minitab, SPSS, R quantile type 6, NumPy method “weibull”Exclusive interpolation, (n + 1)p
R quantile type 1, NumPy method “inverted_cdf”, many service-level and latency reportsNearest rank, ⌈p × n⌉

Percentile of grouped data

When you only have a frequency table, not the raw values, locate the class that contains the pN-th observation and interpolate inside it: Pₖ = L + ((pN − F) / f) × h. Here L is the lower boundary of that class, F the cumulative frequency before it, f its frequency, h the class width and N the total count.

Grouping the 20 exam scores into classes of width 10, each including its lower limit, gives the table below. For the 75th percentile, pN = 0.75 × 20 = 15. The cumulative frequency passes 15 in the 80–90 class, so P75 = 80 + ((15 − 13) / 4) × 10 = 85. That is close to the raw-data answer of 83.5 but not equal, because grouping assumes the values are spread evenly across each class. Use raw data in this calculator whenever you have it; the Frequency Distribution Calculator builds the grouped table from raw data.

The 20 exam scores grouped into classes of width 10
ClassfCumulative f
50–6022
60–7046
70–80713
80–90 (contains the 15th value)417
90–100320

Common use cases

  • Test scores and admissions: convert a raw score into a percentile rank within a class or cohort.
  • Performance monitoring: report P90, P95 or P99 response times, which describe slow requests far better than the mean does.
  • Growth and health data: place a measurement against a reference group, or find the cut-off for the lowest 10%.
  • Pay and pricing: find the 25th, 50th and 75th percentile of salaries or bids to set a band.
  • Checking software output: see at once whether a mismatch with Excel, R or Python is only a different percentile definition.

Assumptions and limitations

  • Percentiles from small samples are rough. With 20 values, anything beyond the 95th percentile is decided by the largest one or two values, whatever method you pick.
  • The methods here work on raw data. Percentiles of grouped (binned) data use the class-boundary formula in the grouped data section above and are only approximate.
  • A percentile rank describes position within this dataset only. To rank a value against a theoretical normal population, convert it to a z-score and use the Z-Score to Percentile Calculator.
  • For Q1 and Q3 by the median-of-halves rule used on TI calculators and in most textbooks, use the Quartile Calculator; that rule is not a percentile formula and can give a different P25 and P75.
  • Percentile ranks are ordinal: the gap between the 50th and 60th rank is usually a much smaller change in the raw value than the gap between the 89th and 99th.

Frequently asked questions

How do you calculate a percentile by hand?

Sort the data, then find a position. With the nearest-rank method, multiply the percentile (as a proportion) by n and round up: the 90th percentile of 20 values is the ⌈0.9 × 20⌉ = 18th value. With interpolation, compute h = (n − 1)p + 1 and move the fractional part of h of the way from one value to the next.

What is the difference between a percentile and a percentile rank?

A percentile is a value: the 90th percentile of a set of scores might be 90.3 marks. A percentile rank is a position between 0 and 100 assigned to a value: a score of 77 might have a percentile rank of 57.5. One starts from a percentage and returns a data value; the other starts from a data value and returns a percentage.

Why does Excel give a different percentile from my textbook?

Excel’s PERCENTILE and PERCENTILE.INC interpolate at position (n − 1)p + 1, while many textbooks use nearest rank or the (n + 1)p position that PERCENTILE.EXC uses. For 20 exam scores the three rules can give 90.3, 90 and 92.7 for the same 90th percentile. None is wrong; they are different conventions, so report which one you used.

What is the difference between PERCENTILE.INC and PERCENTILE.EXC?

PERCENTILE.INC uses position (n − 1)p + 1, so the 0th percentile is the minimum and the 100th is the maximum. PERCENTILE.EXC uses (n + 1)p, which pushes high percentiles a little higher and low ones a little lower, and it returns an error for percentiles below 1 / (n + 1) or above n / (n + 1).

Is the 50th percentile the same as the median?

With either interpolation method, yes: the 50th percentile equals the median for every dataset. With nearest rank and an even number of values, the 50th percentile is the lower of the two middle values, while the median is their average, so the two can differ slightly.

How do you find the 80th percentile of a data set?

Sort the data and compute the position h = (n − 1) × 0.8 + 1. For 20 values, h = 16.2, so the 80th percentile is the 16th value plus 0.2 of the gap to the 17th. With the scores 85 and 88 in those places, P80 = 85 + 0.2 × 3 = 85.6. By nearest rank it is simply the ⌈0.8 × 20⌉ = 16th value, 85.

What is the 95th percentile?

The 95th percentile is the value with 95% of the data at or below it and 5% above. For the 20 exam scores on this page it is 93.15 by the Excel and NumPy method. In web performance and network billing, “P95” means the response time or traffic level that 95% of measurements do not exceed, often found by nearest rank.

How much percentile is 75 marks?

It depends on how everyone else scored, because a percentile rank compares you with a group. Enter all the marks and a value of 75 in this calculator. In the 20 exam scores here, 10 scores are below 75 and none equal it, so 75 marks is at a percentile rank of 10 / 20 × 100 = 50. Scoring 75% on a test and being at the 75th percentile are different things.

How do you convert a percentile rank to a z-score?

Only if the data are roughly normal. Divide the percentile rank by 100 to get a proportion and look it up in the inverse normal table: the 84th percentile is z ≈ 1.0 and the 97.5th is z ≈ 1.96. The Percentile to Z-Score Calculator does this, and the Z-Score to Percentile Calculator goes the other way.

How are ties handled in a percentile rank?

The standard formula PR = (B + 0.5E) / n × 100 counts half of the values tied with x as below it. With 12 values where 4 are below 18 and 3 equal 18, the percentile rank of 18 is (4 + 1.5) / 12 × 100 = 45.83, between the 33.33% strictly below and the 58.33% at or below.

Can a value be at the 100th percentile?

It depends on the definition. Under “% at or below”, the largest value is at 100%. Under the tie-splitting percentile rank formula the maximum of n distinct values gets (n − 0.5) / n × 100, which is below 100, and under “% below” it gets (n − 1) / n × 100. That is why test reports usually top out at the 99th percentile.

Cite this page

Statculator. (2026, November 14). Percentile Calculator. Retrieved November 14, 2026, from https://statculator.com/descriptive-statistics/percentile-calculator/