Percentile vs quartile: how they are related

By Jude Wallis · Published

A percentile is a value with a stated percent of the data at or below it. The quartiles are three particular percentiles: the first quartile is the 25th, the median is the 50th, and the third quartile is the 75th. Every quartile is a percentile, and most percentiles are not quartiles.

Quartiles and percentiles belong to topic 1.7, Summary Statistics for One Quantitative Variable, in Unit 1 of the Fall 2026 AP Statistics course. Finding a percentile of a normal model with a z-score is Unit 2 topic 2.11, The Normal Distribution. This site uses the median-excluded (TI-84) quartile convention and counts values at or below for percentile ranks.

Percentile vs quartile: the short answer

Both mark positions in a distribution, and one is a special case of the other. The kkth percentile is the value at or below which about kk percent of the data fall, and kk can be any percent you choose. The quartiles are the three percentiles that cut the data into four parts of about equal size:

QuartilePercentileAbout this share of the data is at or below it
First quartile, Q1Q_125th25%
Second quartile (the median)50th50%
Third quartile, Q3Q_375th75%

So the difference is one of resolution. Percentiles can locate a value anywhere in the distribution, such as the 90th or the 12th. Quartiles give three fixed landmarks, which are enough to build a five-number summary and a boxplot.

What a percentile tells you

A percentile is a value, in the units of the data. If the 90th percentile of commute times is 48 minutes, then about 90% of commutes took 48 minutes or less. The 90 is the percent and the 48 minutes is the percentile.

You can also run the idea in the other direction. Starting from a value, its percentile rank is the percent of the data at or below it:

percentile rank of x=number of values at or below xn×100\text{percentile rank of } x = \frac{\text{number of values at or below } x}{n} \times 100

This site counts values at or below xx; some sources count only values strictly below, which gives slightly different answers on the same data. On a density curve such as a normal curve, the counting is replaced by area: the percentile rank of xx is the area to the left of xx. The full treatment, including the difference from a percent-correct score, is in what does percentile mean.

What a quartile tells you

The quartiles are found by splitting the sorted data in half at the median. Q1Q_1 is the median of the lower half and Q3Q_3 is the median of the upper half. This site uses the median-excluded (TI-84) rule: when nn is odd, the median itself is left out of both halves. The procedure is in how to find quartiles by hand.

Quartiles do two jobs that general percentiles do not. They define the interquartile range, IQR=Q3−Q1\text{IQR} = Q_3 - Q_1, which is the width of the middle half of the data, and they set the fences of the 1.5 IQR rule for outliers. They are also the edges of the box in a boxplot.

In a small data set, the match between quartiles and percentiles is approximate. For the ten values 52, 58, 60, 64, 65, 69, 73, 79, 86, 91, the lower half is 52, 58, 60, 64, 65, so Q1=60Q_1 = 60, and 3 of the 10 values (30%) are at or below 60. With only ten values, no cut point puts exactly 25% of them at or below it. The larger the data set, the closer the agreement.

The differences side by side

FeaturePercentileQuartile
What it isA value with a chosen percent of the data at or below itOne of three values that split the data into four parts
How many there areOne for each percent you ask aboutThree: Q1Q_1, the median, Q3Q_3
Typical useReporting the relative standing of one valueSummarizing center and spread of a whole distribution
Feeds intoPercentile ranks, growth charts, test score reportsThe IQR, the 1.5 IQR rule, the boxplot
UnitsThe units of the dataThe units of the data
RelationshipThe general ideaThe 25th, 50th, and 75th percentiles

Both are values of the variable. A percent appears only as the label that says which percentile is meant.

When to use which

Use a percentile when the question is about one individual's standing: how a test score, a height, or a race time compares with the rest of the group. A statement such as "her score is at the 83rd percentile" places one value within the distribution.

Use quartiles when the question is about the distribution as a whole. The five-number summary (minimum, Q1Q_1, median, Q3Q_3, maximum) describes center and spread with a median and quartiles that resist outliers, and it is what you report for skewed data. You can compute it with the five-number summary calculator.

When a problem gives a normal model, both come from the same calculation. Any percentile is found by turning the percent into a zz-score and then into a value, and the quartiles are the cases with 25% and 75% of the area to the left, as the second worked example shows.

The classic mix-up and how to avoid it

The most common error is using quartile to mean a group. "She is in the top quartile" is everyday shorthand for being in the highest quarter of the data, above Q3Q_3. In statistics a quartile is a single cut point, a value such as Q3=79Q_3 = 79, and the groups it separates are quarters. When a problem asks for the third quartile, the answer is one number.

The second error is confusing a percentile with a percent score. Scoring 75% on a test means three quarters of the questions were answered correctly. Being at the 75th percentile means about three quarters of the test-takers scored the same or lower. On a hard test, a score of 60% correct can be the 75th percentile.

The third error is reading Q1Q_1 as "25% of the range." Q1Q_1 is the value with about a quarter of the observations at or below it. It is located by counting observations, and it can sit anywhere between the minimum and the median.

Quartiles and a percentile rank from the same data

Twelve test scores, sorted: 52, 55, 58, 61, 64, 67, 70, 73, 77, 81, 86, 94. Find the three quartiles and confirm the percent of scores at or below each. Then find the percentile rank of the score 81.

  1. Find the median. With 12 values it is the average of the 6th and 7th: 67+702=68.5\frac{67 + 70}{2} = 68.5.

  2. Find Q1Q_1, the median of the lower half 52, 55, 58, 61, 64, 67: 58+612=59.5\frac{58 + 61}{2} = 59.5.

  3. Find Q3Q_3, the median of the upper half 70, 73, 77, 81, 86, 94: 77+812=79\frac{77 + 81}{2} = 79.

  4. Count the scores at or below each quartile. At or below 59.5: 52, 55, 58, which is 3 of 12, or 25%. At or below 68.5: 6 of 12, or 50%. At or below 79: 9 of 12, or 75%.

  5. Find the percentile rank of 81. The scores at or below 81 are the first ten in the list, so the rank is 1012×100≈83.3\frac{10}{12} \times 100 \approx 83.3.

  6. Interpret: 81 is at about the 83rd percentile. It is above Q3=79Q_3 = 79, and it is not itself a quartile.

The quartiles are Q1=59.5Q_1 = 59.5, median =68.5= 68.5, and Q3=79Q_3 = 79, with 25%, 50%, and 75% of the scores at or below them. The score 81 has a percentile rank of about 83. The quartiles are the 25th, 50th, and 75th percentiles of this data set, and the 83rd percentile is a percentile that is not a quartile.

Quartiles and the 90th percentile of a normal model

Scores on an exam are modeled by a normal distribution with mean 500 and standard deviation 100. Find Q1Q_1, Q3Q_3, the IQR, and the 90th percentile.

  1. A percentile on a normal curve is the value with the stated area to its left. Find the zz-score for that area, then convert with x=μ+zσx = \mu + z\sigma.

  2. For Q1Q_1 the area to the left is 0.25, and the matching zz-score is about −0.674-0.674. So Q1=500+(−0.674)(100)=432.6Q_1 = 500 + (-0.674)(100) = 432.6.

  3. For Q3Q_3 the area to the left is 0.75. By symmetry the zz-score is about 0.6740.674, so Q3=500+(0.674)(100)=567.4Q_3 = 500 + (0.674)(100) = 567.4.

  4. The interquartile range is 567.4−432.6=134.8567.4 - 432.6 = 134.8, about 135 points. The middle half of the scores spans about 135 points.

  5. For the 90th percentile the area to the left is 0.90, and the matching zz-score is about 1.2821.282. So the 90th percentile is 500+(1.282)(100)=628.2500 + (1.282)(100) = 628.2.

  6. Check the order: 432.6<500<567.4<628.2432.6 < 500 < 567.4 < 628.2, which matches the 25th, 50th, 75th, and 90th percentiles in increasing order.

Q1≈432.6Q_1 \approx 432.6, Q3≈567.4Q_3 \approx 567.4, the IQR is about 135 points, and the 90th percentile is about 628.2. All four come from the same step, converting an area to the left into a zz-score and then into a score. The quartiles are the cases where that area is 0.25 and 0.75.

Frequently asked questions

Is the median a percentile or a quartile?

Both. The median is the 50th percentile, since about half of the data are at or below it, and it is also the second quartile, the middle of the three cut points that divide the data into quarters.

Is the first quartile always exactly the 25th percentile?

In a large data set or on a density curve, yes for practical purposes. In a small data set it is approximate, because the split-the-halves rule for quartiles and the at-or-below count for percentiles can disagree slightly. With ten values, for example, Q1Q_1 can have 30% of the values at or below it.

How many quartiles are there?

Three: Q1Q_1, the median, and Q3Q_3. They divide the sorted data into four quarters. The minimum and the maximum complete the five-number summary, and they are not quartiles.