AP Statistics  /  Unit 1: Exploring One-Variable Data and Collecting Data  /  Topic 1.8
NUM8ERS study notes · Topic 1.8

Graphical Representations of Summary Statisticsfor One Quantitative Variable

Turn five summary numbers into a clear boxplot. Learn what the box, median line, whiskers and outlier points mean—and how to describe the graph without claiming more than it shows.

2026–27 curriculum3 worked examples6 practice questionsSummaries → boxplots → meaning

By the end of this lesson, you should be able to:

  • Connect the five-number summary to the parts of a boxplot.
  • Construct ordinary and modified boxplots on a labeled numerical scale.
  • Use quartiles, whiskers and outlier points to describe center and spread in context.
  • Explain typical mean–median relationships and the limits of a boxplot.

Before you start: Find medians and quartiles and apply the 1.5×IQR⁡1.5\times\operatorname{IQR} rule. Review Topic 1.7 when you need help with the calculations.

First time learning this? Start with the labeled Cedar High plot, then follow the construction checklist and worked examples.

Here to revise? Review the drawing checklist, then try the questions before opening the solutions.

The concept in 60 seconds

A boxplot displays numerical summaries of one quantitative variable. It is also called a box-and-whisker plot. Read its positions on the numerical axis, rather than treating its lines as bars showing frequencies.

The box

Runs from Q1Q_{1} to Q3Q_{3}, showing the middle half of the ordered data. Its length along the numerical axis is the IQR⁡\operatorname{IQR}.

The line inside

Marks the median. It need not sit halfway between the ends of the box.

Whiskers and points

An ordinary plot reaches the minimum and maximum. A modified plot reaches the extreme non-outliers and marks flagged observations separately.

One rule to remember: a modified boxplot’s whisker ends are observed values. Outlier fences are calculated boundaries used to decide which observations are flagged.

A boxplot makes center and spread easy to see. It does not usually show the mean, sample size, every observation or the number of peaks.

Read the Cedar High travel-time boxplot

Return to the same fictional sample of 2020 Cedar High students. Each value is a student’s one-way home-to-school travel time on the Tuesday being studied, rounded to the nearest minute.

Ordered travel times, minutes:
88, 1010, 1212, 1414, 1414, 1515, 1616, 1818, 1818, 1818,
2020, 2222, 2424, 2525, 2626, 2828, 3030, 3232, 3535, 4242

The five-number summary is 88, 14.514.5, 1919, 2727 and 42 min42 \,\mathrm{min}. Here is its boxplot:

Cedar High: locate the five summary numbers

Horizontal boxplot · n=20n = 20 · 1.5×IQR⁡1.5\times\operatorname{IQR} rule

Cedar High: locate the five summary numbersn=20; five-number summary 8, 14.5, 19.0, 27.0, 42; IQR 12.5; fences [-4.25, 45.75]; modified whiskers [8, 42]; flagged observations []. 5 10 15 20 25 30 35 40 45 Recorded one-way travel time (minutes) Minimum 8 Q1 14.5 Median 19 Q3 27 Maximum 42

On a small screen, scroll the plot sideways to read the full scale.

No observations are flagged by the IQR⁡\operatorname{IQR} rule, so the whiskers reach the minimum of 88 and maximum of 42 min42 \,\mathrm{min}. The box spans 14.5–27 min14.5\text{–}27 \,\mathrm{min}, with median 1919. Invented teaching data.

The median line is at 19 min19 \,\mathrm{min}. The box extends from 14.514.5 to 27 min27 \,\mathrm{min}, so IQR⁡=27−14.5=12.5 min\operatorname{IQR}=27-14.5=12.5\,\mathrm{min}. The smallest and largest recorded times are 88 and 42 min42 \,\mathrm{min}, giving a range of 34 min34 \,\mathrm{min}.

Predict: the upper whisker covers a longer interval than the lower whisker. Does that mean the upper part contains more students?

Check your prediction

No. Length along the axis shows how widely values are spread, rather than how many observations are present. Quartiles divide the ordered data into roughly equal quarters; the box covers the middle half. Ties and the quartile convention can affect exact counts in small samples.

Our quartile method takes the median of each half of the sorted data, excluding the overall median from both halves for odd sample sizes. Use the method specified in your question or by your teacher.

The five-number summary and the parts of a boxplot

Minimum · Q1Q_{1} · Median · Q3Q_{3} · Maximum

These are the five-number summary of the full dataset, written in increasing order.

Match a number to its graphical position.
PartWhat to readCommon confusion
First quartile, Q1Q_{1}The smaller-value end of the box; roughly the 25th25^{\mathrm{th}}-percentile position.It is a data position, rather than the lower outlier fence.
Median, Q2Q_{2}The line inside the box; roughly the 50th50^{\mathrm{th}}-percentile position.It is not automatically the mean or geometric midpoint of the box.
Third quartile, Q3Q_{3}The larger-value end of the box; roughly the 75th75^{\mathrm{th}}-percentile position.It is not the maximum.
Whisker endsMinimum and maximum in an ordinary plot; smallest and largest non-outliers in a modified plot.A modified plot’s whisker end need not be an overall extreme.
Separate outlier symbolsFlagged observations beyond the chosen outlier-rule boundaries.The observations remain part of the dataset; they have not been deleted.

Proportions and lengths answer different questions

The intervals minimum–Q1\text{minimum}\text{–}Q_1, Q1–medianQ_1\text{–}\text{median}, median–Q3\text{median}\text{–}Q_3 and Q3–maximumQ_3\text{–}\text{maximum} represent roughly quarters of the ordered data. Those intervals can have very different numerical lengths. A long interval means those values are more spread out.

For Cedar High, the two halves of the box have lengths 19−14.5=4.5 min19-14.5=4.5\,\mathrm{min} and 27−19=8 min27-19=8\,\mathrm{min}. Each corresponds to roughly a quarter of the ordered sample, although their lengths differ.

Modified-plot caution: the outer quarters include any separately plotted outliers. A shortened whisker by itself does not necessarily account for 25%25\% of the observations. With ties, a closed interval bounded by quartiles can include more observations than the simple percentage guide suggests.

Horizontal or vertical: read the numerical axis

These notes use horizontal boxplots, so numerical length is measured left to right. In a vertical boxplot, the same information runs bottom to top. Box thickness perpendicular to that axis is usually a styling choice; it does not encode frequency unless the graph explicitly says it does.

Construct ordinary and modified boxplots

Use a fictional sample of 88 visitor waits from Topic 1.7: 22, 44, 44, 66, 88, 88, 1010 and 22 min22 \,\mathrm{min}. Its five-number summary is 22, 44, 77, 99 and 22 min22 \,\mathrm{min}.

Same 88 waits, different plotting conventions

Common minute scale · box 4–94\text{–}9 · median 77

Ordinary and modified boxplots of the same eight visitor waitsn=8; five-number summary 2, 4.0, 7.0, 9.0, 22; IQR 5.0; fences [-3.5, 16.5]; modified whiskers [2, 10]; flagged observations [22]. Ordinary whiskers are 2 and 22. Modified whiskers are 2 and 10, with an open outlier point at 22. The dashed 16.5 upper fence is a teaching guide, not a whisker endpoint. 0 4 8 12 16 20 24 Recorded visitor wait (minutes) Minimum 2 Maximum 22 Ordinary: whiskers reach the minimum and maximum 0 4 8 12 16 20 24 Recorded visitor wait (minutes) Minimum 2 Fence 16.5 Whisker 10 Outlier 22 Modified: 22 is a separate outlier; whisker ends at 10

On a small screen, scroll the plot sideways to read the full scale.

The dashed upper fence is shown only as a construction guide. The modified plot keeps the 2222-minute observation, displayed separately. Invented teaching data.

Ordinary boxplot: display the five-number summary

  1. Draw a numerical axis with an even scale, variable name and units.
  2. Draw the box from Q1=4Q_{1} = 4 to Q3=9Q_{3} = 9.
  3. Draw the median line at 77.
  4. Extend whiskers to the minimum of 22 and maximum of 2222.

An ordinary plot retains the full endpoints but does not separately identify potential outliers.

Modified boxplot: identify outliers before drawing whiskers

IQR⁡=Q3−Q1\operatorname{IQR}=Q_3-Q_1

Lower fence=Q1−1.5IQR⁡\text{Lower fence}=Q_1-1.5\operatorname{IQR}

Upper fence=Q3+1.5IQR⁡\text{Upper fence}=Q_3+1.5\operatorname{IQR}

  1. Find the quartiles from the full dataset. Here IQR⁡=9−4=5 min\operatorname{IQR}=9-4=5\,\mathrm{min}.
  2. Calculate the fences. Lower=4−1.5(5)=−3.5\text{Lower}=4-1.5(5)=-3.5; upper=9+1.5(5)=16.5 min\text{upper}=9+1.5(5)=16.5\,\mathrm{min}.
  3. Flag observations strictly beyond a fence. Only 2222 is outside these boundaries.
  4. Keep the box and median unchanged. They still come from all 88 observations.
  5. Draw whiskers to actual non-outlier observations. The smallest is 22 and the largest is 1010.
  6. Plot the outlier separately at 2222. Use a clear point, asterisk or other specified symbol.

The upper whisker ends at 1010, not 16.516.5. 10 min10\,\mathrm{min} is the largest recorded wait inside the fence. The 16.516.5-minute fence is a boundary used for classification, not a recorded wait.

A value exactly at a fence is not beyond it and is not flagged by this rule. If several observations are flagged, plot them all; repeated outlier values may overlap unless shown separately. Do not remove outliers and recalculate the quartiles just to construct the modified plot.

Check your drawing or software output

Verify the quartile convention, ordinary/modified plot setting, numerical axis and all plotted observations. Include enough of the axis to show outlier points. A software default may use a different quartile method, so a small discrepancy is a reason to check the method.

The 1.5×IQR⁡1.5\times\operatorname{IQR} rule is the convention used for modified boxplots in this lesson. Topic 1.7 also introduced a separate 22-standard-deviation rule; do not mix its boundaries into this construction.

Mean, median and clues about shape

The mean uses the size of every observation. Extreme values pull it toward their side more strongly than they pull the median. This creates useful typical relationships:

Relatively symmetric

The mean and median are generally close.

Skewed right

The mean is usually greater than the median.

Skewed left

The mean is usually less than the median.

Three examples: the mean is pulled toward the tail

11 dot represents 11 observation · dashed line marks the median · dotted line marks the calculated mean

Three score distributions and their calculated means and mediansSeparate invented nine-score distributions. Symmetric: [11, 12, 13, 13, 14, 15, 15, 16, 17]; mean and median 14. Right tail: [4, 5, 6, 6, 7, 8, 9, 10, 25]; mean 8.8889, median 7. Left tail: [5, 20, 21, 22, 23, 24, 24, 25, 26]; mean 21.1111, median 23. All panels share a 0–30 point scale. 0 5 10 15 20 25 30 Recorded task score (points out of 30) Relatively symmetric · mean 14.00 · median 14 · n = 9 0 5 10 15 20 25 30 Recorded task score (points out of 30) Right-tail example · mean 8.89 · median 7 · n = 9 0 5 10 15 20 25 30 Recorded task score (points out of 30) Left-tail example · mean 21.11 · median 23 · n = 9

On a small screen, scroll the plot sideways to read the full scale.

These dotplots show the underlying values. The added mean lines are explicitly labeled; a standard boxplot does not provide them. Invented teaching data.

What can a boxplot suggest?

A longer extension toward larger values, a median nearer Q1Q_{1}, or high outlier points can suggest right skew. The reverse pattern can suggest left skew. Balanced sides can suggest relative symmetry.

For Cedar High, the upper whisker is 42−27=15 min42-27=15\,\mathrm{min} long, compared with 14.5−8=6.5 min14.5-8=6.5\,\mathrm{min} for the lower whisker. The upper half of the box is longer too. This supports a description of greater extension toward larger travel times. The mean, calculated from the raw data, is 21.35 min21.35 \,\mathrm{min}, above the median of 19 min19 \,\mathrm{min}, consistent with right skew.

Use “suggests” or “consistent with” when appropriate. A boxplot gives shape clues but hides the full distribution. Mean>median\text{Mean}>\text{median} alone does not prove right skew, and mean=median\text{mean}=\text{median} alone does not prove symmetry. Use a dotplot or histogram when the detailed shape matters.

A standard boxplot does not display the mean. The median line supplies the median. Read a mean only if it is provided separately, computed from raw values, or explicitly shown with a labeled additional symbol.

Three worked examples

Example 1: construct Cedar High’s boxplot

Prompt: Construct a modified boxplot for the 2020 travel times and explain whether 42 min42 \,\mathrm{min} should be a separate outlier point.

  1. Summaries: minimum 88, Q1Q_{1} 14.514.5, median 1919, Q3Q_{3} 2727, maximum 4242.
  2. IQR⁡\operatorname{IQR}: 27−14.5=12.5 min27-14.5=12.5\,\mathrm{min}.
  3. Fences: 14.5−1.5(12.5)=−4.2514.5-1.5(12.5)=-4.25; 27+1.5(12.5)=45.75 min27+1.5(12.5)=45.75\,\mathrm{min}.
  4. Outliers: every recorded time is within the fences, including 4242.
  5. Drawing: box 14.5–2714.5\text{–}27, median line 1919, whiskers to 88 and 4242, no separate outlier symbols.

Model explanation: The 4242-minute journey is not flagged by the 1.5×IQR⁡1.5\times\operatorname{IQR} rule because it is below the upper fence of 45.75 min45.75 \,\mathrm{min}. The ordinary and modified plots therefore have the same appearance for these data.

Being the largest observation alone does not make a value an outlier.

Example 2: locate whiskers when a value is flagged

Prompt: For the 88 visitor waits, give the modified plot’s whisker endpoints and the full data range.

IQR⁡=5 min\operatorname{IQR} = 5 \,\mathrm{min} and the fences are −3.5-3.5 and 16.5 min16.5 \,\mathrm{min}. The 2222-minute wait is flagged, so the whiskers extend to 22 and 10 min10 \,\mathrm{min}. Plot 2222 as a separate point.

Model response: The modified boxplot has whiskers at 22 and 10 min10 \,\mathrm{min} and a high outlier point at 22 min22 \,\mathrm{min}. The full data range is 22−2=20 min22-2=20\,\mathrm{min}. The 88-minute distance between the whisker ends is not the full range because it excludes the flagged observation.

The maximum remains 2222 in the five-number summary. A modified plot changes how it is shown, rather than changing the underlying dataset.

Example 3: read a boxplot and describe it in context

A separate fictional sample of 88 students has reading-time summaries minimum 1212, Q1Q_{1} 1818, median 2222, Q3Q_{3} 3030 and maximum 45 min45 \,\mathrm{min}.

Read the summaries: 88 recorded reading times

Horizontal boxplot · n=8n = 8 · 1.5×IQR⁡1.5\times\operatorname{IQR} rule

Read the summaries: eight recorded reading timesn=8; five-number summary 12, 18.0, 22.0, 30.0, 45; IQR 12.0; fences [0.0, 48.0]; modified whiskers [12, 45]; flagged observations []. 10 15 20 25 30 35 40 45 50 Recorded reading time (minutes) Minimum 12 Q1 18 Median 22 Q3 30 Maximum 45

On a small screen, scroll the plot sideways to read the full scale.

The box spans 18–30 min18\text{–}30 \,\mathrm{min} and the median is 22 min22 \,\mathrm{min}. The IQR⁡\operatorname{IQR} rule flags no observations, so the whiskers reach 1212 and 45 min45 \,\mathrm{min}. Invented teaching data.

Prompt: Describe center, spread and the visible shape clue. Can you calculate the exact mean from this graph alone?

  1. Center: median=22 min\text{median} = 22 \,\mathrm{min}.
  2. Spread: IQR⁡=30−18=12 min\operatorname{IQR}=30-18=12\,\mathrm{min}; full range=45−12=33 min\text{full range}=45-12=33\,\mathrm{min}.
  3. Shape clue: the graph extends farther toward larger reading times. That suggests right skew; it does not reveal the number of peaks.
  4. Limit: the exact mean cannot generally be recovered from the boxplot alone.

Model response: For these 8 sampled students\text{these }8\text{ sampled students}, the median reading time is 22 min22 \,\mathrm{min} and the middle half spans about 18–30 min18\text{–}30 \,\mathrm{min}, an IQR⁡\operatorname{IQR} of 12 min12 \,\mathrm{min}. Recorded times extend from 1212 to 45 min45 \,\mathrm{min}. The longer extension toward larger values suggests right skew. The standard boxplot does not provide an exact mean.

The figure has no flagged observations under the 1.5×IQR⁡1.5\times\operatorname{IQR} rule: its fences are 00 and 48 min48 \,\mathrm{min}. “No points drawn” only means no outliers were identified by the plot’s stated rule, rather than proving there are no unusual observations under every possible criterion.

Explain a boxplot in context—and know its limits

A useful description names the variable, recorded group, numerical evidence and units. Start with the median and IQR⁡\operatorname{IQR}, then add full extent, flagged observations and cautious shape clues when relevant.

Sentence starter: “For [recorded group], the median [variable] is [value and units], and the middle half extends from [Q1Q_{1}] to [Q3Q_{3}], giving an IQR⁡\operatorname{IQR} of [difference and units]. The graph also shows [supported feature].”

What a standard boxplot does and does not tell you

Match your claim to the available information.
You can usually read or calculateYou generally cannot recover from the plot alone
Median, quartiles and IQR⁡\operatorname{IQR}.The exact mean and standard deviation.
Full endpoints if all extreme points are shown; full range using whiskers and any outlier symbols.Every individual value or exact frequencies within a numerical interval.
Flagged outliers under the stated plotting rule.Whether a flagged point is an error, or why it occurred.
Shape clues from unequal lengths and isolated points.The number of peaks, specific internal gaps, clusters or a guarantee of normality.
Approximate relative positions in quarters of the ordered data.Sample size unless labeled, or an exact percentage at an arbitrary cutoff.
Identical boxes can hide different observations

Dotplot above each ordinary boxplot · common point scale · same five-number summary

Two sets with identical boxplots and different meansSet A values [0, 1, 2, 2, 4, 5, 5, 6, 8, 8, 9, 10] have mean 5. Set B values [0, 2, 2, 2, 5, 5, 5, 8, 8, 8, 10, 10] have mean 65/12, approximately 5.42. Both five-number summaries are 0,2,5,8,10. Dotplots show repeated values with stacks; both boxplots are identical. 0 2 4 6 8 10 Recorded task score (points) Set A · n = 12 · calculated mean = 5.00 points 0 2 4 6 8 10 Recorded task score (points) Set B · n = 12 · calculated mean ≈ 5.42 points

On a small screen, scroll the plot sideways to read the full scale.

The values differ while the minimum, quartiles, median and maximum stay the same. The boxes alone do not reveal the difference in means. Invented teaching data.

These 22 invented sets share the five-number summary 00, 22, 55, 88 and 10 points10 \,\mathrm{points}, yet their observations and means differ. Their ordinary boxplots therefore look identical. A graph of the full observations preserves information that the summary omits.

See the 22 sets and check the means

Set A: 00, 11, 22, 22, 44, 55, 55, 66, 88, 88, 99, 1010. Total=60\text{Total} = 60; mean=5 points\text{mean} = 5 \,\mathrm{points}.
Set B: 00, 22, 22, 22, 55, 55, 55, 88, 88, 88, 1010, 1010. Total=65\text{Total} = 65; mean≈5.42 points\text{mean} \approx 5.42 \,\mathrm{points}.

For both sets, Q1=2+22=2Q_1=\frac{2+2}{2}=2, median=5+52=5\text{median}=\frac{5+5}{2}=5, and Q3=8+82=8Q_3=\frac{8+8}{2}=8. The extreme values are also the same.

Can five summary numbers always give a complete modified plot?

Not when an endpoint is flagged and the remaining observations are unknown. The summary lets you locate the box and median and test the minimum and maximum against fences. It does not necessarily give the extreme non-outlier values or all separately plotted outliers.

For example, a summary of 00, 1010, 1212, 1414 and 3030 has IQR⁡\operatorname{IQR} 44 and fences 44 and 2020. The minimum 00 and maximum 3030 are flagged, but the modified whisker endpoints require additional data. Do not invent observations at 44 or 2020.

If a task supplies a graph rather than exact numbers, respect its scale and use “about” where readings are estimates. Boxplot summaries alone do not establish a cause or justify generalizing from a sample to every member of a population.

Common mistakes and how to fix them

Check the meaning of each line and number.
MistakeWhy it failsBetter approach
Ending a whisker at the fenceA fence is a calculated boundary, not necessarily an observation.Use the extreme observed value within the fences.
Calling the line inside the box the meanThat line marks the median.Read a mean only when separately supplied or explicitly labeled.
Saying a longer box segment contains more observationsLength encodes numerical spread.Use the approximate quarter/half guide and account for ties.
Using whisker-to-whisker distance as the full range in a modified plotSeparate extreme points can be outside the whiskers.Find the overall minimum and maximum, including flagged points.
Removing outliers before calculating quartiles for the plotIt changes the dataset and the summary.Use all observations, then display flagged values separately.
Treating a balanced boxplot as proof of symmetry or normalityThe summary hides the detailed pattern.Describe the visible balance cautiously and inspect a fuller graph if needed.
Mixing the 22-standard-deviation rule into an IQR⁡\operatorname{IQR}-based boxplotThe rules can flag different observations.Use the plotting convention stated in the task.

Six practice questions with hints and solutions

Use the median-of-halves quartile method from this lesson unless a question provides the summary. All data are invented teaching examples.

1. Build a modified boxplot from raw values

1010 students’ recorded reading times are 66, 88, 99, 1010, 1212, 1212, 1414, 1515, 1616 and 28 min28 \,\mathrm{min}. Find the five-number summary, fences, whisker endpoints and any outlier points.

Need a hint?

The 2 middle values2\text{ middle values} are both 1212. Take the median of each 5-value half5\text{-value half}. Then calculate 1.5×IQR⁡1.5\times\operatorname{IQR}.

Show the solution

Minimum=6\text{Minimum} = 6, Q1=9Q_{1} = 9, median=12\text{median} = 12, Q3=15Q_{3} = 15, maximum=28 min\text{maximum} = 28 \,\mathrm{min}. IQR⁡=15−9=6 min\operatorname{IQR}=15-9=6\,\mathrm{min}. Lower fence=9−1.5(6)=0\text{Lower fence}=9-1.5(6)=0 and 15+1.5(6)=24 min15+1.5(6)=24\,\mathrm{min}.

Only 2828 is flagged. Draw the box from 99 to 1515, the median line at 1212, whiskers to 66 and 1616, and a separate point at 2828 on a labeled minute scale.

2. Interpret unequal box segments

A reading-time boxplot has Q1=18Q_{1} = 18, median=22\text{median} = 22 and Q3=30 minQ_{3} = 30 \,\mathrm{min}. Find its IQR⁡\operatorname{IQR} and the length of each half of the box. Does the longer half prove that more students are in that part?

Need a hint?

Subtract the endpoints of each interval. Then distinguish numerical length from proportion.

Show the solution

IQR⁡=30−18=12 min\operatorname{IQR}=30-18=12\,\mathrm{min}. Lower box segment=22−18=4 min\text{Lower box segment}=22-18=4\,\mathrm{min}; upper segment=30−22=8 min\text{upper segment}=30-22=8\,\mathrm{min}. Each represents roughly a quarter of the ordered data. The upper segment’s greater length shows greater spread over that quarter, rather than proving a greater number of students. Exact closed-interval counts depend on ties and the quartile convention.

3. Range versus whisker span

A modified wait-time boxplot has whiskers at 22 and 10 min10 \,\mathrm{min}, a box from 44 to 99, median 77, and 1 outlier point1\text{ outlier point} at 2222. Calculate the full range and IQR⁡\operatorname{IQR}. What is the dataset’s maximum?

Need a hint?

Keep the separate point when identifying the full extremes.

Show the solution

Maximum=22 min\text{Maximum} = 22 \,\mathrm{min}; minimum=2\text{minimum} = 2. Full range=22−2=20 min\text{Full range}=22-2=20\,\mathrm{min}. IQR⁡=9−4=5 min\operatorname{IQR}=9-4=5\,\mathrm{min}. The whisker span 10−2=8 min10-2=8\,\mathrm{min} excludes the outlier and is not the full range.

4. Mean–median reasoning

A distribution of recorded scores has a long tail toward smaller values. Its median is 23 points23 \,\mathrm{points}. Would you usually expect its mean to be above or below 2323? Can you find the exact mean from a standard boxplot alone?

Need a hint?

The mean is pulled toward the tail. A typical relationship is different from an exact calculation.

Show the solution

This is a left-skewed pattern, so the mean would usually be below 23 points23 \,\mathrm{points}. The lower scores pull the mean toward the left tail. A standard boxplot alone generally does not give the exact mean; raw values or an additional mean summary would be needed.

5. What is missing from a summary?

A dataset has five-number summary 00, 1010, 1212, 1414 and 3030. Can you determine a complete modified boxplot using only these five numbers? Explain what you can draw and what additional information you need.

Need a hint?

Calculate the fences, test the minimum and maximum, and ask whether you know the next observation inside each fence.

Show the solution

IQR⁡=4\operatorname{IQR} = 4, so the fences are 44 and 2020. You can draw the box from 1010 to 1414 and the median line at 1212. The extreme observations 00 and 3030 are flagged.

You need raw data or additional information identifying the smallest and largest non-outliers and all flagged observations. There might be other outliers between an extreme and its fence. Ending whiskers at 44 and 2020 would invent endpoints not known to be observed.

6. Diagnose a student’s drawing

For the 88 visitor waits, a student draws the box at 4–94\text{–}9, places the median line at 6.56.5, extends the upper whisker to 16.516.5 and erases the 2222-minute wait. Identify three errors and correct them.

Need a hint?

The box midpoint is not necessarily the median. A fence is not a whisker endpoint. Flagged observations remain in the display.

Show the solution
  1. The median is 7 min7 \,\mathrm{min}, found from 6+82\frac{6+8}{2}, rather than the geometric box midpoint 6.56.5.
  2. The upper whisker must end at the largest non-outlier, 10 min10 \,\mathrm{min}, rather than the 16.516.5-minute fence.
  3. Keep 22 min22 \,\mathrm{min} as a separate outlier point. Do not erase the observation.

The lower whisker is at 2 min2 \,\mathrm{min} and the box remains 4–9 min4\text{–}9 \,\mathrm{min}.

Quick revision checklist

Before drawing or interpreting a boxplot.
CheckWhat to do
Order and methodSort raw values; use the stated quartile convention.
Five-number summaryMinimum, Q1Q_{1}, median, Q3Q_{3}, maximum—keep the full dataset.
Plot typeOrdinary: whiskers to extremes. Modified: apply the stated rule and show flagged observations separately.
IQR⁡\operatorname{IQR} and fencesIQR⁡=Q3−Q1\operatorname{IQR}=Q_3-Q_1; fences Q1−1.5IQR⁡Q_1-1.5\operatorname{IQR} and Q3+1.5IQR⁡Q_3+1.5\operatorname{IQR}.
Whisker endpointsUse actual observations; equality at a fence is not beyond it.
Axis and labelsEven scale, variable name, units and all outlier positions.
InterpretationMedian for center; IQR⁡\operatorname{IQR} for middle-half spread; include outliers when finding the full range.
Shape and meanUse cautious shape clues. Symmetric: mean≈median\text{mean}\approx\text{median}; right skew: usually mean>median\text{mean}>\text{median}; left skew: usually mean<median\text{mean}<\text{median}.
LimitsA standard boxplot generally does not give the mean, standard deviation, nn, individual values, peaks or gaps.

Before submitting: did I place the median correctly, distinguish fences from observations, retain outlier points and explain the graph with context and units?

Final understanding check

1212 sampled deliveries have recorded waits of 55, 66, 77, 88, 99, 1010, 1010, 1111, 1212, 1313, 1414 and 30 min30 \,\mathrm{min}.

  1. Find the five-number summary and IQR⁡\operatorname{IQR} using the median-of-halves method.
  2. Find the 1.5×IQR⁡1.5\times\operatorname{IQR} fences and construct a modified boxplot, labeling the axis and outlier points.
  3. Calculate the full range and describe center and spread in context.
  4. The sample mean, calculated from the raw values, is 11.25 min11.25 \,\mathrm{min}. How does it compare with the median, and why is this relationship reasonable here? Does the boxplot alone reveal the exact cause of the long wait?
Show the full solution and completed plot

1. Minimum=5\text{Minimum} = 5; Q1=7+82=7.5Q_1=\frac{7+8}{2}=7.5; median=10+102=10\text{median}=\frac{10+10}{2}=10; Q3=12+132=12.5Q_3=\frac{12+13}{2}=12.5; maximum=30 min\text{maximum} = 30 \,\mathrm{min}. IQR⁡=12.5−7.5=5 min\operatorname{IQR}=12.5-7.5=5\,\mathrm{min}.

2. Lower fence=7.5−1.5(5)=0\text{Lower fence}=7.5-1.5(5)=0 and 12.5+1.5(5)=20 min12.5+1.5(5)=20\,\mathrm{min}. Only 3030 is flagged. Draw the box from 7.57.5 to 12.512.5, median line at 1010, whiskers to 55 and 1414, and an outlier point at 3030.

Completed modified boxplot: delivery waits

n=12n = 12 · box 7.5–12.57.5\text{–}12.5 · median 1010 · fences 00 and 2020

Final-check modified delivery wait boxplotn=12; five-number summary 5, 7.5, 10.0, 12.5, 30; IQR 5.0; fences [0.0, 20.0]; modified whiskers [5, 14]; flagged observations [30]. 0 5 10 15 20 25 30 35 Recorded delivery wait (minutes) Whisker ends 5 and 14 Outlier 30

On a small screen, scroll the plot sideways to read the full scale.

The 3030-minute wait is outside the upper fence. Its separate point remains part of the data; the full range is 25 min25 \,\mathrm{min}. Invented teaching data.

3. Full range=30−5=25 min\text{Full range}=30-5=25\,\mathrm{min}. For these 1212 sampled deliveries, the middle recorded wait is 10 min10 \,\mathrm{min}, and the middle half spans about 7.5–12.5 min7.5\text{–}12.5 \,\mathrm{min}, an IQR⁡\operatorname{IQR} of 5 min5 \,\mathrm{min}. The 3030-minute wait is flagged by the 1.5×IQR⁡1.5\times\operatorname{IQR} rule.

4. The mean of 11.25 min11.25 \,\mathrm{min} is greater than the median of 10 min10 \,\mathrm{min}. This is consistent with the high isolated observation pulling the mean toward larger values. The boxplot does not establish the reason for that wait or prove a cause.

Ready to move on? You should be able to construct and read a modified boxplot, distinguish summaries from observations, explain the typical mean–median relationships and state what a boxplot leaves out.

Continue learning

Review summary-statistic calculations · Review modified-plot construction · Back to the lesson overview