What is a decile?
Deciles split an ordered data set into ten equal parts. There are nine deciles, labelled D1 through D9. The first decile (D1) is the value below which 10% of the data fall, D5 is the median (50%), and D9 marks the 90% point. Deciles are widely used in statistics, education (test-score banding), economics (income distribution) and finance to describe how a value ranks within a population.
How to use this calculator
Enter your numbers separated by commas or spaces, then choose which decile you want (k from 1 to 9). The calculator sorts the values, computes the rank position with the formula \(L = k(n+1)/10\), and returns the decile value — interpolating between two data points when the position is not a whole number.
The formula explained
For \(n\) sorted values, the position of the k-th decile is \(L = k(n+1)/10\). If \(L\) is an integer, the decile is simply the value at that rank. If \(L\) falls between ranks (say 3.4), the decile lies 40% of the way from the 3rd value toward the 4th: $$D = x_3 + 0.4\,(x_4 - x_3).$$ Positions at or below 1 use the smallest value; positions at or above \(n\) use the largest.
Worked example
Take the data 2, 4, 6, 8, 10, 12, 14, 16, 18, 20 (\(n = 10\)) and find D3. The position is $$L = 3 \times (10 + 1) / 10 = 3.3.$$ So D3 lies 30% of the way from the 3rd value (6) to the 4th (8): $$D_3 = 6 + 0.3 \times (8 - 6) = 6.6.$$
More Worked Examples
Each example below follows the same three steps: sort the data, compute the rank position \(P = \dfrac{k\,(n+1)}{10}\), then split it into the integer part \(L = \lfloor P \rfloor\) and the fraction \(f = P - L\) and interpolate between the \(L\)th and \((L+1)\)th sorted values.
Example 1 — a whole-number rank (n = 19, k = 5)
Raw input: 2, 4, 5, 7, 8, 10, 11, 13, 14, 16, 18, 19, 21, 23, 25, 26, 28, 30, 32
Sorted list (already ascending), with ranks 1–19:
2, 4, 5, 7, 8, 10, 11, 13, 14, 16, 18, 19, 21, 23, 25, 26, 28, 30, 32 → \(n = 19\)
Rank position:
$$P = \frac{5\,(19+1)}{10} = \frac{5 \times 20}{10} = \frac{100}{10} = 10$$Interpolation: \(L = \lfloor 10 \rfloor = 10\) and \(f = 10 - 10 = 0\), so the fractional term vanishes:
$$D_5 = x_{(10)} + 0 \times \bigl(x_{(11)} - x_{(10)}\bigr) = 16 + 0 \times (18 - 16) = 16$$Final value: \(D_5 = \) 16. Because \(P\) landed exactly on an integer, no interpolation was needed and \(D_5\) is simply the 10th sorted observation — which for \(n = 19\) is also the median.
Example 2 — small, unsorted n with repeated values (n = 7)
Raw input: 12, 5, 9, 5, 20, 9, 14
Sort step: 5, 5, 9, 9, 12, 14, 20 → \(n = 7\), so ranks run 1 through 7 and \(n + 1 = 8\). The repeated 5s occupy ranks 1 and 2; the repeated 9s occupy ranks 3 and 4. Ties are kept as separate observations — they are not merged.
D1:
$$P = \frac{1 \times 8}{10} = 0.8, \qquad L = \lfloor 0.8 \rfloor = 0, \qquad f = 0.8$$Rank 0 does not exist (there is no \(x_{(0)}\) below the smallest value), so the target position lies outside the data and the result is clamped to the minimum:
$$D_1 = x_{(1)} = 5$$D9:
$$P = \frac{9 \times 8}{10} = 7.2, \qquad L = \lfloor 7.2 \rfloor = 7, \qquad f = 0.2$$Interpolation would require \(x_{(8)}\), but the list stops at rank 7, so the result is clamped to the maximum:
$$D_9 = x_{(7)} = 20$$D5 for the same data lands cleanly inside the list:
$$P = \frac{5 \times 8}{10} = 4, \qquad L = 4,\; f = 0 \quad\Rightarrow\quad D_5 = x_{(4)} = 9$$Final values: \(D_1 = 5\), \(D_5 = \) 9, \(D_9 = 20\). With only 7 observations, \(D_1\) and \(D_9\) collapse onto the smallest and largest values; a data set needs \(n \ge 9\) for \(P\) to fall at or inside rank 1 for \(k = 1\), and \(n \ge 9\) likewise for \(k = 9\).
Example 3 — 25 exam scores, find D8
Raw input (scores in the order they were recorded):
72, 85, 58, 90, 66, 41, 78, 93, 62, 70, 99, 55, 80, 68, 48, 96, 74, 60, 82, 52, 91, 63, 76, 65, 71
Sorted list (ranks 1–25):
41, 48, 52, 55, 58, 60, 62, 63, 65, 66, 68, 70, 71, 72, 74, 76, 78, 80, 82, 85, 90, 91, 93, 96, 99 → \(n = 25\)
Rank position for the 8th decile:
$$P = \frac{8\,(25+1)}{10} = \frac{8 \times 26}{10} = \frac{208}{10} = 20.8$$So \(L = \lfloor 20.8 \rfloor = 20\) and \(f = 20.8 - 20 = 0.8\). The 20th sorted score is \(x_{(20)} = 85\) and the 21st is \(x_{(21)} = 90\).
Interpolation:
$$D_8 = x_{(20)} + 0.8\,\bigl(x_{(21)} - x_{(20)}\bigr) = 85 + 0.8\,(90 - 85) = 85 + 0.8 \times 5 = 85 + 4 = 89$$Final value: \(D_8 = \) 89 points. Under this tool's convention, a score of 89 marks the cut-off at which about 80% of the 25 recorded scores fall at or below it. Note that 89 is not one of the observed scores — interpolation produces a value between the 20th and 21st observations, which is normal and expected whenever \(f \neq 0\).
Decile, Percentile and Quartile Equivalents
Deciles, percentiles and quantile fractions are three notations for the same cut-points: \(D_k\) is the \(10k\)th percentile, i.e. the quantile at fraction \(k/10\). The last column evaluates this tool's rank formula \(L = \dfrac{k(n+1)}{10}\) for a sample of \(n = 20\) (so \(n + 1 = 21\)) to show the concrete position in the sorted list.
| Decile | Equivalent percentile | Quantile fraction | Quartile equivalent | Rank \(P = k(n+1)/10\) for \(n = 20\) | Interpolation needed? |
|---|---|---|---|---|---|
| D1 | 10th percentile | 0.1 | — | \(1 \times 21/10 = 2.1\) | Yes: \(x_{(2)} + 0.1\,(x_{(3)} - x_{(2)})\) |
| D2 | 20th percentile | 0.2 | — | \(2 \times 21/10 = 4.2\) | Yes: \(x_{(4)} + 0.2\,(x_{(5)} - x_{(4)})\) |
| D3 | 30th percentile | 0.3 | — | \(3 \times 21/10 = 6.3\) | Yes: \(x_{(6)} + 0.3\,(x_{(7)} - x_{(6)})\) |
| D4 | 40th percentile | 0.4 | — | \(4 \times 21/10 = 8.4\) | Yes: \(x_{(8)} + 0.4\,(x_{(9)} - x_{(8)})\) |
| D5 | 50th percentile | 0.5 | \(Q_2\) = median (exact, for any \(n\)) | \(5 \times 21/10 = 10.5\) | Yes: \(x_{(10)} + 0.5\,(x_{(11)} - x_{(10)})\) — the usual even-\(n\) median |
| D6 | 60th percentile | 0.6 | — | \(6 \times 21/10 = 12.6\) | Yes: \(x_{(12)} + 0.6\,(x_{(13)} - x_{(12)})\) |
| D7 | 70th percentile | 0.7 | — | \(7 \times 21/10 = 14.7\) | Yes: \(x_{(14)} + 0.7\,(x_{(15)} - x_{(14)})\) |
| D8 | 80th percentile | 0.8 | — | \(8 \times 21/10 = 16.8\) | Yes: \(x_{(16)} + 0.8\,(x_{(17)} - x_{(16)})\) |
| D9 | 90th percentile | 0.9 | — | \(9 \times 21/10 = 18.9\) | Yes: \(x_{(18)} + 0.9\,(x_{(19)} - x_{(18)})\) |
Note on quartiles: only the middle cut-point is shared between the two systems — \(D_5 = P_{50} = Q_2\) = the median. No decile coincides with \(Q_1\) or \(Q_3\). The first quartile sits at the 25th percentile, which would be "\(D_{2.5}\)", and the third quartile at the 75th percentile, i.e. "\(D_{7.5}\)" — and \(k = 2.5\) and \(k = 7.5\) are not deciles, since \(k\) takes only the integer values 1 through 9. \(Q_1\) therefore falls between \(D_2\) and \(D_3\), and \(Q_3\) between \(D_7\) and \(D_8\). To obtain \(Q_1\) or \(Q_3\) with the same \((n+1)\) convention, use the rank \(P = \dfrac{j(n+1)}{4}\) for \(j = 1, 3\) instead.
Note on the boundaries: \(k = 0\) and \(k = 10\) are not deciles either — they would correspond to the 0th and 100th percentiles, i.e. the minimum and maximum of the data set. The nine deciles create ten intervals, so with \(n\) observations each interdecile interval contains about \(n/10\) of them.
FAQ
What is the difference between a decile and a percentile? Deciles divide data into 10 parts; percentiles into 100. Dk equals the (10k)-th percentile, so D3 is the 30th percentile.
Why do I get a decimal that isn't in my data? When the computed rank position is not a whole number, the calculator interpolates between the two nearest values, so the result can be a number not present in your original list.
Does the order of my input matter? No — the calculator sorts the data for you, so you can paste numbers in any order.