Describing a Boxplot's Shape and Skew

Use Unequal Spacing as Evidence, Not Proof

In Describing a Dotplot With Gaps and Clusters, you practiced describing features a graph shows directly. A boxplot preserves less detail than a dotplot, but its quartiles and whiskers can still suggest whether a distribution is more spread out on one side of its center than the other. This asymmetry can provide evidence of skew.

In Reading a Boxplot’s Quartiles and Spread, you learned to locate the quartiles, median, and whisker endpoints. To assess shape, compare the numerical lengths of the sections on either side of the median. A box with a longer upper half, for example, shows more spread from the median to \(Q_3\) than from \(Q_1\) to the median. A longer right whisker can also suggest that values extend farther above the box than below it.

Definition: A boxplot suggests right skew when its upper-side sections generally extend farther than its lower-side sections, and suggests left skew when its lower-side sections generally extend farther. Roughly balanced sections are consistent with approximate symmetry. These visual patterns are clues, not definitive proof of a distribution’s shape.

The word “suggests” matters. A boxplot summarizes selected locations; it does not show how observations are arranged between those locations. As covered in What a Boxplot Cannot Show, it cannot reveal every concentration, gap, or peak. A histogram or dotplot of the same data may show details that help clarify the shape.

Compare the Four Distances

For a basic boxplot, compare four distances along the number line: from the minimum to \(Q_1\), from \(Q_1\) to the median, from the median to \(Q_3\), and from \(Q_3\) to the maximum. These are not supposed to be equal. Instead, look for a consistent imbalance: are the sections on one side of the median generally longer?

\begin{aligned} \text{lower whisker length} &= Q_1-\text{minimum}\\ \text{lower half of box} &= \text{median}-Q_1\\ \text{upper half of box} &= Q_3-\text{median}\\ \text{upper whisker length} &= \text{maximum}-Q_3 \end{aligned}

The intervals between the minimum, \(Q_1\), median, \(Q_3\), and maximum each span about one-fourth of the ordered observations. Their numerical lengths can differ greatly. A long section means that the values in that part of the distribution cover a wider range; it does not mean that the section contains more observations.

The two halves of the box are particularly useful to compare because they meet at the median. If the upper half of the box is much longer than the lower half, the values from the median to \(Q_3\) are more spread out than the values from \(Q_1\) to the median. This is evidence consistent with right skew. If the lower half is much longer, the box suggests left skew. If they are similar in length, the middle half is roughly balanced around the median, though that alone does not establish that the full distribution is symmetric.

Whiskers add evidence about the spread outside the box. A longer right whisker suggests more extension above \(Q_3\); a longer left whisker suggests more extension below \(Q_1\). Consider the box and both whiskers together. If one side is longer in both the box and whiskers, the evidence for asymmetry is more consistent than if only one section differs.

Key distinction: In a basic boxplot, whiskers extend to the minimum and maximum. In a modified boxplot, whiskers generally extend only to the most extreme observations that are not flagged by the 1.5 IQR rule; flagged observations are plotted separately. Thus, visible whisker-to-quartile sections in a modified boxplot exclude those outliers and need not span about one-fourth of the observations.

That distinction affects how you interpret length. A short whisker on a modified boxplot does not necessarily mean there is little spread in that direction: separately plotted outliers may lie beyond it. Check whether the display is modified and whether individual points appear outside the whiskers. You can then describe the whisker pattern and any plotted points separately, without treating the whisker endpoints as the full minimum and maximum.

Worked Examples

Worked Example: Longer Spread Above the Median

A fictional boxplot summarizes the number of minutes students at a school spend traveling to school. It is a basic boxplot with minimum 12, \(Q_1=16\), median 20, \(Q_3=29\), and maximum 47. Describe what its spacing suggests about shape.

State. The variable is students’ travel time to school, measured in minutes. The boxplot shows the quartiles and the full range.

Plan. Calculate the lengths of the two box halves and two whiskers. Compare the sections below and above the median, then describe the resulting evidence cautiously.

Do. The lower whisker is \(16-12=4\) minutes. The lower half of the box is \(20-16=4\) minutes. The upper half of the box is \(29-20=9\) minutes. The upper whisker is \(47-29=18\) minutes. Thus, both sections above the median are longer than their counterparts below it: 9 is greater than 4 in the box, and 18 is greater than 4 in the whiskers.

\begin{aligned} \text{lower whisker} &= 16-12=4\\ \text{lower box half} &= 20-16=4\\ \text{upper box half} &= 29-20=9\\ \text{upper whisker} &= 47-29=18 \end{aligned}

Conclude. The boxplot suggests right skew in students’ travel times: the spread above the median is greater than the spread below it, both within the box and in the whisker. This describes the pattern of the summary; it does not identify why some travel times are much longer.

Worked Example: More Spread Below the Median

A fictional basic boxplot summarizes the number of hours each plant in a greenhouse takes to reach a specified growth stage. Its five-number summary is minimum 8, \(Q_1=12\), median 23, \(Q_3=27\), and maximum 31 hours. What does the spacing suggest?

State. The quantitative variable is time to reach the growth stage, measured in hours. This basic boxplot’s whiskers reach the minimum and maximum.

Plan. Find the four section lengths, comparing the spread below the median with the spread above it.

Do. The lower whisker is \(12-8=4\) hours, and the lower half of the box is \(23-12=11\) hours. The upper half of the box is \(27-23=4\) hours, and the upper whisker is \(31-27=4\) hours. The lower half of the box is substantially longer than the upper half, while the whiskers have equal lengths.

\begin{aligned} \text{lower whisker} &= 12-8=4\\ \text{lower box half} &= 23-12=11\\ \text{upper box half} &= 27-23=4\\ \text{upper whisker} &= 31-27=4 \end{aligned}

Conclude. The boxplot suggests some left-side asymmetry in the plants’ growth times because the lower half of the box is longer. The equal whisker lengths do not show the same imbalance, so it would be too strong to claim that every section indicates left skew. A careful description reports the mixed evidence rather than treating one feature as conclusive.

Worked Example: A Modified Boxplot With a High Outlier

A fictional modified boxplot summarizes the number of kilometers per week that members of a walking club record. It shows \(Q_1=40\), median 45, \(Q_3=52\), whisker endpoints at 34 and 61, and a separate point at 90. Describe the visible asymmetry and the point carefully.

State. The variable is distance walked per week, measured in kilometers. Because this is a modified boxplot, the whisker endpoints may not be the minimum and maximum.

Plan. Compare the displayed sections, then use the IQR and the 1.5 IQR rule from Applying the 1.5 IQR Rule for Outliers to check whether the point at 90 is beyond the upper fence. Keep the whisker comparison separate from the separately plotted point.

Do. The lower half of the box is \(45-40=5\) kilometers, and the upper half is \(52-45=7\) kilometers. The lower whisker is \(40-34=6\) kilometers, and the upper whisker is \(61-52=9\) kilometers. Both upper-side sections are somewhat longer. Also, \(\text{IQR}=52-40=12\) kilometers, so the upper fence is \(52+1.5(12)=70\) kilometers. Since \(90>70\), the point at 90 is above the upper fence and is flagged by the rule.

\begin{aligned} \text{lower box half} &= 45-40=5\\ \text{upper box half} &= 52-45=7\\ \text{lower whisker} &= 40-34=6\\ \text{upper whisker} &= 61-52=9\\ \text{IQR} &= 52-40=12\\ \text{upper fence} &= 52+1.5(12)=70 \end{aligned}

Conclude. The quartile and whisker sections suggest modest right-side asymmetry in weekly walking distances. The point at 90 kilometers is separately plotted and lies above the upper fence; it is not part of the upper whisker. Describe it as a value flagged by the 1.5 IQR rule, not as proof of a particular cause or error.

Phrase the Inference Precisely

A strong answer connects the visual evidence to the shape claim. Instead of writing only “the boxplot is right-skewed,” name the sections that are longer and say that they suggest right skew. For example: “The upper half of the box and the right whisker are longer than the corresponding sections below the median, suggesting that the distribution of weekly travel times is right-skewed.” The statement identifies the evidence, the direction, the variable, and the context.

If the evidence is mixed, say so. A longer upper whisker paired with a longer lower half of the box is not a consistent indication in one direction. You might report that the whisker extends farther to the right, while the box halves are similar or point the other way. Avoid forcing a single shape label when the sections do not support it.

If all four distances are roughly balanced, you can say the boxplot is consistent with approximate symmetry. Do not say it proves symmetry: different distributions can have similar quartiles and whiskers. Likewise, unequal lengths do not reveal the exact shape of a tail or where observations are concentrated. A histogram or dotplot can provide more detail, as discussed in Matching Histograms to Boxplots.

Key takeaway: Compare both halves of the box and both whiskers. Consistently longer sections on the right suggest right skew; consistently longer sections on the left suggest left skew. Describe the evidence and use “suggests” or “is consistent with,” especially when the sections disagree or the plot is modified.

Common Mistakes and AP Exam Tips

Naming skew without citing evidence. A full-credit description points to the unequal box halves or whiskers and identifies which side extends farther.
Calling a longer section more populated. The quartile intervals each cover about one-fourth of the ordered observations. Their unequal lengths indicate different numerical spread, not different counts.
Assuming each visible modified-boxplot section covers one-fourth of the data. Modified whiskers can stop at the most extreme non-outliers, with flagged observations plotted separately. Do not treat their endpoints as the full minimum and maximum.
Using one unequal section as certain proof of skew. Compare the box and whiskers together. If the evidence is mixed, describe the mismatch instead of overstating a direction.
Forgetting context and units. State which variable and group the boxplot summarizes, and include units when discussing numerical lengths.
Claiming that a boxplot shows peaks or gaps. It does not display the detailed arrangement of observations. Use a histogram or dotplot to investigate those features.
Calling a separately plotted point an error. A point beyond a fence is flagged by a rule; that alone does not establish that it was measured incorrectly or explain why it is unusual.

Check Your Understanding

For each question, focus on what the boxplot’s spacing supports and phrase any shape inference cautiously.

A basic boxplot of a river’s daily water levels has a lower box half of 3 centimeters, an upper box half of 8 centimeters, a left whisker of 4 centimeters, and a right whisker of 13 centimeters. What shape does the spacing suggest, and what evidence supports your answer?
A boxplot has approximately equal whisker lengths, but its lower half of the box is much longer than its upper half. Why should your description mention mixed evidence rather than make a strong overall claim?
In a modified boxplot, the upper whisker is short and two points are plotted above it. Why would it be incorrect to treat the whisker endpoint as the maximum?
What does a longer upper half of the box tell you about the values between the median and \(Q_3\)? What does it not tell you about the number of observations in that interval?
Write a careful one-sentence description for a boxplot whose four sections are roughly balanced. Include the variable’s context and avoid claiming that the boxplot proves symmetry.