Quick Answer: How Do You Read a Box Plot?
Find the median line inside the box. Next locate Q1 and Q3 at the box edges. The distance from Q1 to Q3 is the interquartile range, or IQR, which describes the spread of the middle 50% of the data. Then inspect the whiskers and any points beyond them. Finally, compare the two sides of the box and whiskers for asymmetry, while remembering that whisker rules can vary.
Box Plot Anatomy at a Glance
The numerical axis tells you the data values. Along that axis, the box runs from the first quartile to the third quartile, and the line inside it marks the median. In a common modified box plot, whiskers extend to the most extreme observed values that remain inside the outlier fences.
| Part | What to look for | What it tells you | What not to assume |
|---|---|---|---|
| Median | Line inside the box | Central location by the 50th percentile | It is not automatically the mean |
| Q1 and Q3 | Two ends of the box | Bounds of the central quartile interval | Quartile algorithms are identical in every program |
| IQR | Length of the box on the numeric axis | Spread of the middle 50% | A larger IQR always means a larger standard deviation |
| Whiskers | Lines extending beyond the box | Extent of non-flagged observations under the chosen rule | They always equal the minimum and maximum |
| Potential outliers | Points beyond the whiskers | Observations flagged by the plot's rule | They are data-entry errors or should be deleted |
| Asymmetry | Unequal box halves or whiskers | Possible uneven spread across the distribution | One long whisker proves population skewness |
What Is a Box Plot?
A box plot, also called a box-and-whisker plot, summarizes the location and spread of a numerical distribution using quartiles, a median, whiskers, and sometimes individually plotted potential outliers.
Box plots are compact. That makes them useful when several groups must fit on the same chart. The tradeoff is detail: a box plot does not show every observation, every gap, or every cluster. If you need a deeper foundation on the graph itself, see the site's box plot guide. For the broader topic this belongs to, return to Descriptive Statistics.
How to Read a Box Plot in 7 Steps
A repeatable reading order
1. Read the numerical scale
Start with the axis, not the box. A value of 30 could mean 30 seconds, 30 years, $30, or a score of 30. Also check whether several box plots use the same axis limits. Visual length only supports a fair comparison when the scales match.
2. Find the median
The median is the line inside the box. It divides the ordered observations into lower and upper halves in the usual descriptive sense. If Group A has a median of 60 and Group B has a median of 72, Group B has the higher median. That statement is different from saying it has the higher mean.
3. Identify Q1 and Q3
Q1 is the first quartile and Q3 is the third quartile. They form the two boundaries of the box. In a percentile interpretation, Q1 is around the 25th percentile and Q3 is around the 75th percentile. Exact finite-sample quartiles can differ slightly across calculation methods, so small datasets may produce different endpoints in different software.
4. Compare the IQR
The interquartile range is Q3 minus Q1. It measures the spread of the central 50% of the observations. A larger IQR means the middle half is more spread out. A smaller IQR means it is more concentrated. Learn the calculation in more detail in the interquartile range guide, or calculate it with the IQR Calculator.
5. Inspect the whiskers
Whiskers show how far observations extend beyond the box under the chart's chosen convention. In a common modified box plot, each whisker stops at the most extreme observed value that is still inside the 1.5 × IQR fence. A whisker does not have to end exactly on the calculated fence.
6. Look for potential outliers
Values beyond the whiskers may be drawn as dots, circles, or other markers. Treat them as observations flagged by the rule, not as automatic mistakes. A flagged point can be genuine variation, an unusual case, a measurement problem, or a data-entry issue. Investigation comes before deletion. The site's outliers guide explains that decision in more depth.
7. Assess asymmetry and compare groups
Look at the median's position inside the box, the two halves of the box, and both whiskers together. If several features stretch farther toward higher values, the pattern can be consistent with right-side asymmetry. If they stretch farther toward lower values, it can be consistent with left-side asymmetry. Use those as clues, not proof of the population's shape.
What Does Each Part of a Box Plot Mean?
Q1: the first quartile
Q1 marks the lower edge of the box. It is a lower quartile boundary of the ordered distribution under the quartile method being used. The interval from Q1 to the median contains one quarter of the distribution in the percentile sense.
Median: the line inside the box
The median is Q2, the 50th percentile. It does not need to sit in the geometric center of the box. If the distance from Q1 to the median is much shorter than the distance from the median to Q3, the upper half of the box is more spread out.
Q3: the third quartile
Q3 marks the upper edge of the box. Together, Q1 and Q3 define the middle half of the distribution. If you want to review the full five-number framework, see the five-number summary.
IQR: the middle spread
The IQR is more resistant to extreme values than the full range because it depends on the middle part of the data rather than the smallest and largest observations.
Whiskers: observed endpoints under a rule
The whiskers are easy to misread. In the common modified convention used throughout the worked examples below, the lower whisker ends at the smallest observed value at or above the lower fence. The upper whisker ends at the largest observed value at or below the upper fence. Other software or chart authors may use minimum/maximum or percentile-based whiskers instead.
Potential outliers: individually plotted observations
Points outside the whiskers are observations that the chosen rule has flagged. Calling them potential outliers is more careful than calling them errors. A box plot tells you where the observation falls relative to the quartiles. It does not tell you why the observation is unusual.
Why Whiskers Do Not Always Mean Minimum and Maximum
Do not automatically label the two whisker tips as the dataset minimum and maximum. In a modified box plot, extreme observations beyond the 1.5 × IQR fences are plotted separately, so the whiskers stop at the most extreme non-flagged observations.
This distinction also separates a fence from a whisker. A fence is a calculated threshold. A whisker endpoint is usually an actual observed value. If an upper fence is 42 but the largest observation inside it is 39, the upper whisker ends at 39, not 42.
Current Matplotlib documentation uses 1.5 × IQR by default and allows other whisker settings, including percentile-based endpoints. R's base boxplot function also exposes its whisker range as a configurable argument. When a published chart matters, check its legend or software documentation rather than guessing the convention.
How to Identify Potential Outliers on a Box Plot
For a common modified box plot, calculate the IQR first. Then multiply the IQR by 1.5 and subtract that amount from Q1 for the lower fence, while adding it to Q3 for the upper fence. Observations beyond those fences are flagged.
Worked Example: A Box Plot With a High-Side Potential Outlier
Use the sorted data: 8, 10, 12, 14, 16, 18, 20, 22, 24, 26, 60. For this example, quartiles are calculated with the median-of-halves method, excluding the overall median from the two halves.
The median is 18. The lower half is 8, 10, 12, 14, 16, giving Q1 = 12. The upper half is 20, 22, 24, 26, 60, giving Q3 = 24. Therefore IQR = 24 − 12 = 12.
Lower fence = 12 − 1.5(12) = −6. Upper fence = 24 + 1.5(12) = 42. The value 60 lies beyond 42, so it is flagged as a potential outlier. The upper whisker ends at 26, the largest observed value still inside the fence.
How to Read Skewness From a Box Plot
A box plot can show visual clues of asymmetry, but it cannot reconstruct the full distribution. For a pattern that looks right-skewed, you may see the median closer to Q1, a longer median-to-Q3 section, a longer upper whisker, or high-side flagged values. A left-skewed pattern may show the reverse.
Use several features together. One long whisker can reflect a few widely spaced observations without proving that the population is skewed. A roughly centered median and similar whiskers can look symmetric, but symmetry in a box plot does not prove normality. For shape measures, see skewness and kurtosis. For a normality-focused diagnostic, a Q-Q plot gives different information.
How to Compare Two Box Plots
Compare the same features in the same order: median, IQR, whiskers, potential outliers, then overall asymmetry. If the plots do not share the same numerical scale, stop and resolve that before comparing lengths.
Worked Example: Comparing Delivery Times
Suppose Group A delivery times are 22, 24, 25, 26, 28, 30, 32, 34, 36, 38, 40. Group B times are 20, 22, 24, 26, 28, 35, 42, 44, 46, 48, 80.
Using the same median-of-halves method, Group A has Q1 = 25, median = 30, Q3 = 36, so IQR = 11. Group B has Q1 = 24, median = 35, Q3 = 46, so IQR = 22. Group B has the higher median and more spread in the middle 50%.
For Group B, the upper fence is 46 + 1.5(22) = 79, so 80 is flagged and the upper whisker ends at 48. The box plot shows a descriptive difference. It does not tell you whether the groups differ statistically significantly.
- Center: compare medians.
- Middle spread: compare IQRs, shown by box length on the numeric axis.
- Tails: compare whisker lengths under the same convention.
- Potential outliers: note flagged points and which side they occur on.
- Asymmetry: use the median position, box halves, and whiskers together.
- Inference: do not treat box overlap or non-overlap as a significance test.
If you want to build the comparison interactively, use the site's Box Plot Comparison Tool.
Worked Example: Reading a Straightforward Box Plot
Take the sorted observations 10, 15, 20, 25, 28, 30, 35, 40, 45, 50, 55. With the median-of-halves method, the median is 30, Q1 is 20, and Q3 is 45. The IQR is 25.
The lower fence is 20 − 1.5(25) = −17.5. The upper fence is 45 + 1.5(25) = 82.5. Every observation lies within those fences, so there are no flagged points. The whiskers extend to 10 and 55.
Reading the plot, you would say the middle 50% spans 20 to 45, the median is 30, and the upper half of the box is longer than the lower half. That last feature suggests more spread above the median inside the box, but it is not enough by itself to declare a particular population shape.
How the Five-Number Summary Relates to a Box Plot
The traditional five-number summary is minimum, Q1, median, Q3, and maximum. A skeletal box plot can draw whiskers directly to the minimum and maximum. A modified box plot adds an outlier rule, which means the plotted whisker endpoints can differ from the minimum and maximum.
The five-number summary describes five summary values. A modified box plot may use Q1, median, and Q3 from that summary while drawing its whiskers to adjacent non-flagged observations rather than to the raw minimum and maximum.
You can practice these summaries with the Five-Number Summary Visualizer or the Box Plot Calculator.
Box Plot vs. Histogram
A box plot and a histogram answer different questions. The box plot is compact and makes group comparisons easy. The histogram uses more space but shows distribution shape in greater detail.
| Question | Box plot | Histogram |
|---|---|---|
| Where is the median? | Shown directly | Usually not shown directly |
| How wide is the middle 50%? | Shown by the IQR | Must be inferred or calculated separately |
| Are there multiple peaks? | Usually hidden | Can be visible |
| Are there gaps or clusters? | Usually hidden | Can be visible |
| Can many groups be compared compactly? | Very useful | Often becomes crowded |
See the histogram guide or build one with the Histogram Maker.
What a Box Plot Cannot Tell You
A box plot hides detail by design. Two datasets can share similar quartiles and whiskers while having very different raw shapes. A standard box plot may not reveal multimodality, gaps, dense clusters, exact sample size, or the number and spacing of observations inside the whiskers.
It also does not normally display the mean. Some software can overlay a mean marker, but you should not infer a mean that is not drawn. Likewise, a symmetric-looking box plot does not establish normality, and ordinary box overlap does not determine statistical significance.
Sample size deserves its own check. Two boxes can look nearly identical even if one summarizes 20 observations and another summarizes 20,000. Some variable-width box plots encode sample size in the box width, but ordinary box width is often just styling. Read the legend before assigning meaning to it.
Common Box Plot Interpretation Mistakes
Calling whiskers min and max by default
Check the convention first. Modified plots often stop whiskers at the most extreme non-flagged observations.
Confusing fences with whiskers
Fences are calculated thresholds. Whisker tips are usually observed data values.
Calling every flagged point an error
A flagged value may be valid. Investigate it before deciding how to handle it.
Treating median as mean
The standard box plot shows the median. It does not automatically show the arithmetic mean.
Using box thickness as spread
Spread is measured along the numerical axis. The perpendicular width is usually decorative unless the chart says otherwise.
Equating symmetry with normality
A box plot hides too much shape information to prove that a distribution is normal.
Using one whisker as proof of skew
Look for a pattern across the median position, box halves, whiskers, and flagged points.
Inferring significance from overlap
Box overlap is descriptive. Formal inference depends on the design, sample size, variability, and analysis method.
Why Different Programs Can Produce Slightly Different Box Plots
Box plots are not governed by one universal recipe. Programs can differ in how they calculate sample quartiles, how they define whiskers, and whether they use 1.5 × IQR, percentiles, or the full range. Those differences matter most in small datasets where a change in Q1 or Q3 can also change the IQR and the outlier fences.
For the worked examples on this page, quartiles use the median-of-halves method and whiskers use the common 1.5 × IQR modified convention. If your textbook or software reports a slightly different Q1 or Q3 for the same small dataset, first compare the quartile rule before assuming one result is wrong.
Frequently Asked Questions
Key Takeaways
- Read the numerical scale before interpreting the box.
- The line inside the box is the median, not automatically the mean.
- Q1 and Q3 form the box, and IQR = Q3 − Q1 measures middle spread.
- Whisker meaning depends on the plotting convention.
- In a common modified box plot, fences flag potential outliers while whiskers end at actual observations inside the fences.
- Potential outliers should be investigated rather than automatically deleted.
- Use multiple visual clues when discussing asymmetry or skewness.
- Compare groups by median, IQR, whiskers, and flagged points, but do not infer statistical significance from box overlap alone.
- A box plot hides detailed density, clusters, gaps, and often sample size.
Sources and Statistical References
The interpretation rules in this guide were checked against university teaching resources and current statistical software documentation.
Related Resources on Statistics Fundamentals
Try the Box Plot Generator with a short dataset, then compare your reading against the Box Plot Calculator. Watching the quartiles and whiskers change makes the reading order easier to remember.