Graphical Representations of Summary Statistics - AP Statistics

SECTION 1

Historical Context & Motivation

Long before digital dashboards and statistical software, researchers wrestled with the problem of communicating large amounts of numerical data in a form that human perception could rapidly comprehend. Raw tables of numbers — means, medians, quartiles, standard deviations — provided precise numerical answers, but they lacked the capacity to convey distributional shape, identify outliers, or compare groups at a glance. The drive to represent summary statistics graphically grew out of this tension between precision and perception, and it has shaped the practice of exploratory data analysis ever since.

1786 • Playfair's Bar and Line Charts

William Playfair published The Commercial and Political Atlas, introducing bar charts and line graphs to represent economic data — among the first systematic uses of graphical displays for quantitative summaries.

1900 • Karl Pearson's Histograms

Karl Pearson formalized the histogram as a tool to display frequency distributions, enabling analysts to see center, spread, and skewness in a single image.

1977 • Tukey's Exploratory Data Analysis

John Tukey published Exploratory Data Analysis, introducing the boxplot (box-and-whisker plot) and stem-and-leaf display — graphical tools built directly from summary statistics like the five-number summary.

1999 • AP Statistics Curriculum Established

The College Board's AP Statistics course formalized the expectation that students not only compute summary statistics but also represent them graphically and interpret those representations in context.

The central question this lesson addresses is: how do we take numerical summaries — the mean, median, quartiles, range, interquartile range, and standard deviation — and embed them in visual displays that communicate distributional characteristics efficiently and accurately? Mastering this skill is essential because the AP Statistics exam frequently asks you to construct, interpret, and compare graphical displays, and to connect what you see in a graph to the underlying numerical summaries.

SECTION 2

Core Principles & Definitions

Before constructing any graphical representation, it is essential to understand the summary statistics that underpin the displays. A summary statistic is a single number that captures one aspect of a distribution — its center, its variability, or its position within a dataset. Graphical representations map these numbers onto visual elements such as position, length, area, and spacing, enabling the viewer to perceive distributional features that raw numbers cannot convey as readily.

Measures of Center

The mean (arithmetic average) and median (middle value of ordered data) describe where the 'typical' value lies. On a graph, these are represented as marked positions along the number line or within a display.

Measures of Spread

The range, interquartile range (IQR), and standard deviation quantify how much observations vary. Graphically, spread translates to the width or extent of a display.

Five-Number Summary

The minimum, first quartile (Q₁), median (Q₂), third quartile (Q₃), and maximum together form the five-number summary — the foundation of the boxplot.

Shape Descriptors

Symmetry, skewness (left or right), modality (unimodal, bimodal), and the presence of outliers describe the overall form of the distribution and are visible only through graphical displays.

✦

KEY TAKEAWAY

Think of summary statistics as the coordinates of a building, and graphical displays as the architectural blueprint. The coordinates (numbers) are precise and essential, but the blueprint (graph) lets you see whether the structure is balanced, lopsided, or has unusual features — information that a list of coordinates alone cannot convey. Each graphical tool is designed to make certain summary statistics visually salient.

SECTION 3

Visual Explanation — The Boxplot

The boxplot (also called a box-and-whisker plot) is the most direct graphical representation of summary statistics, encoding the entire five-number summary and identifying potential outliers in a single compact display. The diagram below shows a horizontal boxplot for the dataset: 12, 15, 18, 20, 22, 24, 25, 27, 30, 35, 48. The five-number summary is: Min = 12, Q₁ = 18, Median = 24, Q₃ = 30, Max = 48. Using the 1.5 × IQR rule, 48 is flagged as an outlier because it exceeds Q₃ + 1.5 × IQR = 30 + 1.5(12) = 48, which is right at the boundary — for pedagogical clarity, we treat it as an outlier in this example.

A horizontal boxplot showing the five-number summary. The cyan line marks the median; the violet box spans Q₁ to Q₃ (the IQR); whiskers extend to the most extreme non-outlier values; and the red circle denotes an outlier beyond 1.5 × IQR from the nearest quartile.

Notice how each structural element of the boxplot maps directly onto a summary statistic. The left edge of the box is Q₁, the right edge is Q₃, and the width of the box encodes the IQR — the spread of the middle 50% of the data. The whiskers extend outward to the smallest and largest observations that fall within 1.5 × IQR of the quartiles, while any point beyond that threshold appears as an individually plotted marker. This construction means that a boxplot allows immediate comparison of center (median position), spread (box width and whisker length), and symmetry (relative lengths of the two halves of the box and whiskers), making it one of the most information-dense single-variable displays available.

SECTION 4

Mathematical Framework

Constructing graphical representations of summary statistics requires computing the statistics themselves. Here we formalize the key formulas that feed directly into boxplots, dotplots with marked statistics, and other displays. A solid grasp of these calculations ensures you can both construct graphs by hand and verify calculator output on the AP exam.

SAMPLE MEAN

x̄ = (1/n) × Σxᵢ for i = 1, 2, …, n

where x̄ is the sample mean, n is the number of observations, and xᵢ represents each individual data value. The mean is the balance point of the distribution and is sensitive to outliers.

INTERQUARTILE RANGE

IQR = Q₃ − Q₁

where Q₁ is the 25th percentile and Q₃ is the 75th percentile. The IQR measures the spread of the central 50% of data and defines the width of the box in a boxplot.

OUTLIER BOUNDARIES

Lower fence = Q₁ − 1.5 × IQR Upper fence = Q₃ + 1.5 × IQR

Any observation below the lower fence or above the upper fence is classified as a potential outlier and plotted individually beyond the whiskers of a boxplot.

SAMPLE STANDARD DEVIATION

s = √[ Σ(xᵢ − x̄)² / (n − 1) ]

The standard deviation s measures the typical distance of observations from the mean. While not directly encoded in a boxplot, it governs the spread visible in dotplots and histograms, and is central to the interpretation of distributions described by mean ± s.

When you construct a boxplot, you first compute the five-number summary (Min, Q₁, Median, Q₃, Max), then calculate the fences to determine where whiskers terminate and which observations are outliers. For displays such as dotplots with annotated summary statistics, you additionally compute x̄ and s to mark on the number line. The key insight is that every visual element in a well-constructed statistical graph corresponds to a specific computed quantity — there is no guesswork.

SECTION 5

Detailed Breakdown of Display Types

Although the boxplot is the most explicit graphical representation of summary statistics, several other displays encode or reveal summary statistics in different ways. Understanding the strengths and limitations of each display type is critical for selecting the right tool in an AP Statistics free-response question and for interpreting computer output correctly.

Two displays of the same dataset. The dotplot preserves individual values and annotates center with marked lines for mean and median, while the histogram reveals distributional shape through bar heights but obscures individual data points.

Display Type	Summary Statistics Shown	Best Used When
Boxplot	Min, Q₁, Median, Q₃, Max, outliers (via 1.5 × IQR rule)	Comparing distributions across groups; identifying outliers and skewness
Dotplot	Individual values; can annotate mean, median, and spread on the axis	Small datasets (n < 30); showing every observation with summary annotations
Histogram	Shape (symmetry, skewness, modality); approximate center and spread from bar patterns	Larger datasets; assessing distributional shape before choosing summary statistics
Stem-and-Leaf Plot	Exact values; shape; allows quick computation of median and quartiles	Small to moderate datasets where preserving individual values matters; back-to-back comparisons

Common graphical displays and the summary statistics they encode or reveal

When choosing a display on the AP exam, consider what the question asks you to communicate. If you need to compare the medians and spreads of two or more groups, side-by-side boxplots are the most efficient choice. If the question asks about the shape of a single distribution or whether the data are approximately normal, a histogram or stemplot is more informative because boxplots cannot reveal bimodality or gaps. Dotplots are ideal when you want to show individual values alongside annotated summary measures for small samples.

SECTION 6

Worked Example

A researcher records the number of hours per week that 15 college students spend studying. The ordered dataset is: 4, 6, 8, 10, 12, 14, 15, 16, 18, 20, 22, 24, 28, 32, 45. Construct a modified boxplot and describe the distribution.

Constructing a Modified Boxplot

1
Step 1 — Compute the Five-Number Summary

With n = 15 observations, the median is the 8th value: Median = 16. The lower half consists of the first 7 values (4, 6, 8, 10, 12, 14, 15), so Q₁ is the 4th value: Q₁ = 10. The upper half consists of the last 7 values (18, 20, 22, 24, 28, 32, 45), so Q₃ is the 4th value from that group: Q₃ = 24. Min = 4, Max = 45.

Five-Number Summary: 4, 10, 16, 24, 45
2
Step 2 — Compute the IQR and Fences

IQR = Q₃ − Q₁ = 24 − 10 = 14. The lower fence is Q₁ − 1.5 × IQR = 10 − 21 = −11. The upper fence is Q₃ + 1.5 × IQR = 24 + 21 = 45.

IQR = 14; Lower fence = −11; Upper fence = 45
3
Step 3 — Identify Outliers

Check each observation against the fences. No observation falls below −11, and the maximum value of 45 is exactly at the upper fence. Since observations must exceed (not equal) the fence to be classified as outliers under the standard convention, 45 is not an outlier in this dataset. All observations lie within the fences.

No outliers identified
4
Step 4 — Draw the Boxplot

Draw a number line from 0 to 50. Place the left edge of the box at Q₁ = 10 and the right edge at Q₃ = 24. Draw a vertical line inside the box at the median = 16. Extend the lower whisker from Q₁ left to the minimum value of 4, and extend the upper whisker from Q₃ right to the maximum value of 45 (the largest non-outlier value).

Boxplot: whisker 4–10 | box 10–24 with median at 16 | whisker 24–45
5
Step 5 — Describe the Distribution

The distribution of weekly study hours is skewed right because the upper whisker (from 24 to 45, length 21) is substantially longer than the lower whisker (from 4 to 10, length 6), and the median (16) is closer to Q₁ than to Q₃. The center is approximately 16 hours, and the IQR of 14 hours indicates moderate variability. The right skew suggests that a few students study substantially more than the typical student.

Right-skewed, center ≈ 16 hours, IQR = 14 hours, no outliers

SECTION 7

Strengths, Limitations, and Comparisons

Each graphical representation of summary statistics involves trade-offs. No single display communicates every aspect of a distribution perfectly, and recognizing the strengths and limitations of each display is a recurring theme on the AP exam. The table below synthesizes the key advantages and drawbacks of the most common graphical displays you will encounter.

Display	Strengths	Limitations
Boxplot	Concisely displays five-number summary; excellent for side-by-side group comparisons; identifies outliers automatically via the 1.5 × IQR rule	Cannot reveal bimodality, gaps, or clusters within the distribution; does not show the mean or standard deviation; obscures sample size
Histogram	Reveals distributional shape including modality, gaps, and clusters; handles large datasets well; allows visual estimation of center and spread	Bin width choice affects appearance; does not display exact values; comparing multiple groups requires separate panels
Dotplot	Shows every individual observation; intuitive for small datasets; easy to annotate with mean, median, and spread markers	Becomes cluttered with large datasets; repeated values require stacking that can be hard to read
Stemplot	Preserves exact data values while showing shape; back-to-back version allows direct two-group comparison; easy to compute summary statistics directly from the display	Impractical for large datasets or data with many decimal places; difficult to construct for non-numeric or very spread-out data

Strengths and limitations of common graphical displays for summary statistics

✦

KEY TAKEAWAY

Choosing a graphical display is analogous to selecting the right diagnostic tool in medicine: an X-ray (boxplot) gives a quick structural overview, while an MRI (histogram) reveals internal detail. A blood pressure reading (numerical summary alone) tells you a number, but the imaging (graph) shows where and why. On the AP exam, always match the display to what the question asks you to communicate.

SECTION 8

Connection to Advanced Theory

The graphical tools covered in this lesson form the foundation of exploratory data analysis (EDA), which serves as the essential preliminary step before any formal inference. When you move into later units of AP Statistics — particularly those dealing with confidence intervals and hypothesis tests — the habit of first graphing your data becomes critical for verifying assumptions such as approximate normality, checking for outliers that may invalidate certain procedures, and assessing whether the mean or median is a more appropriate measure of center.

This Lesson (EDA)	Advanced Application
Boxplot reveals skewness and outliers	Informs choice between t-procedures (mean-based) vs. non-parametric methods (median-based) for inference
Histogram assesses approximate normality	Validates the normality condition required for constructing confidence intervals and conducting z/t-tests
Side-by-side boxplots compare group distributions	Motivates two-sample t-tests and ANOVA by visualizing whether group centers appear significantly different
Mean vs. median comparison indicates skewness	Guides the choice of resistant vs. non-resistant summary statistics in regression diagnostics and residual analysis

How graphical EDA connects to inference procedures in later AP Statistics units

Beyond the AP course, the ideas in this lesson extend into advanced statistical visualization methods such as violin plots (which combine boxplots with kernel density estimates), quantile-quantile plots (which compare observed quantiles against theoretical distributions), and interactive dashboards used in data science. The core principle remains the same: every visual element should correspond to a meaningful statistical quantity, and the viewer should be able to read summary statistics directly from the graph without ambiguity.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A boxplot of test scores shows that the median line is very close to Q₁, the upper whisker is much longer than the lower whisker, and there are two individually plotted points above the upper whisker. Which of the following best describes this distribution?

A. Approximately symmetric with two outliers
B. Skewed left with two outliers
C. Skewed right with two outliers
D. Bimodal and skewed right
E. Uniform with two outliers

Show answer (C)

The correct answer is C. When the median is close to Q₁ and the upper whisker is much longer than the lower whisker, the distribution is skewed right — the data stretch out toward higher values. The two individually plotted points above the upper whisker are outliers identified by the 1.5 × IQR rule. A is incorrect because a symmetric distribution would have the median near the center of the box with roughly equal whisker lengths. B is incorrect because left skew would produce a longer lower whisker and a median near Q₃. D is incorrect because boxplots cannot reveal bimodality — that requires a histogram or dotplot. E is incorrect because a uniform distribution would appear roughly symmetric with equal whisker lengths.

Question 2: BASIC CALCULATION

Consider the dataset: 3, 7, 8, 12, 14, 18, 21, 25, 30. What is the IQR, and what are the outlier fences for a modified boxplot?

A. IQR = 13; Lower fence = −12; Upper fence = 42.5
B. IQR = 15.5; Lower fence = −15.75; Upper fence = 46.25
C. IQR = 17; Lower fence = −18; Upper fence = 48.5
D. IQR = 27; Lower fence = −33.5; Upper fence = 63.5
E. IQR = 15.5; Lower fence = 0; Upper fence = 46

Show answer (B)

The correct answer is B. With n = 9 ordered values {3, 7, 8, 12, 14, 18, 21, 25, 30}, the median is the 5th value, which is 14. Using the standard method (TI-84 / AP Statistics convention), the lower half consists of the four values below the median: {3, 7, 8, 12}. Q₁ is the median of this lower half: Q₁ = (7 + 8)/2 = 7.5. The upper half consists of the four values above the median: {18, 21, 25, 30}. Q₃ is the median of this upper half: Q₃ = (21 + 25)/2 = 23. The IQR = Q₃ − Q₁ = 23 − 7.5 = 15.5. For a modified boxplot, the outlier fences are calculated as: Lower fence = Q₁ − 1.5 × IQR = 7.5 − 1.5(15.5) = 7.5 − 23.25 = −15.75. Upper fence = Q₃ + 1.5 × IQR = 23 + 1.5(15.5) = 23 + 23.25 = 46.25. Since no data points fall below −15.75 or above 46.25, this dataset has no outliers. Choice A is incorrect because it uses IQR = 13, which results from an error in the quartile calculation. Choice C is incorrect because it uses IQR = 17, which does not result from any standard quartile method for this dataset. Choice D is incorrect because it uses the range (27) instead of the IQR. Choice E is incorrect because it uses rounded fence values that do not match the correct computation.

Question 3: INTERMEDIATE

Two classes took the same exam. Class A's boxplot shows: Min = 55, Q₁ = 68, Median = 75, Q₃ = 82, Max = 95. Class B's boxplot shows: lower whisker endpoint = 40, Q₁ = 60, Median = 74, Q₃ = 88, Max = 98, with an outlier at 32. Which of the following comparisons is best supported by the boxplots?

A. Class A has a higher center and greater variability than Class B.
B. Class A and Class B have similar centers, but Class B has greater variability.
C. Class B has a higher center and less variability than Class A.
D. Both classes have approximately the same center and the same variability.
E. Class A is more skewed than Class B.

Show answer (B)

The correct answer is B. The medians are nearly identical (75 vs. 74), indicating similar centers. However, Class B's IQR = 88 − 60 = 28 is much larger than Class A's IQR = 82 − 68 = 14, showing Class B has substantially greater variability in the middle 50% of scores. Class B also has a wider whisker span and an outlier at 32, further emphasizing its greater spread. A is incorrect because although Class A's median is only 1 point higher (essentially the same), Class A has less variability, not more. C is incorrect because Class B does not have a higher center and certainly does not have less variability. D is incorrect because while the centers are similar, the variabilities are very different (IQR of 14 vs. 28). E is incorrect because Class A's boxplot (median closer to center of box, similar whisker lengths) actually appears more symmetric than Class B's, which has a low outlier pulling the distribution toward the left.

Question 4: APPLIED

4 pts

A pharmaceutical researcher collects the following data on patient recovery times (in days) for a new treatment: 5, 7, 8, 9, 10, 10, 11, 12, 13, 14, 15, 17, 22, 28, 42. (a) Calculate the five-number summary and determine whether any outliers exist using the 1.5 × IQR rule. (b) Construct a modified boxplot based on your calculations. Clearly label all components. (c) The researcher's colleague argues that a histogram would be more appropriate than a boxplot for this dataset. Provide one reason supporting the colleague's argument and one reason supporting the use of a boxplot.

Show answer

Rubric (4 points): Part (a) — Five-number summary and outlier check (2 points) • (1 point) Correct five-number summary: Min = 5, Q₁ = 9, Median = 12, Q₃ = 17, Max = 42. With n = 15, the median is the 8th value = 12. Q₁ is the median of the lower 7 values (5, 7, 8, 9, 10, 10, 11), which is the 4th value = 9. Q₃ is the median of the upper 7 values (13, 14, 15, 17, 22, 28, 42), which is the 4th value = 17. • (1 point) Correct outlier identification: IQR = 17 − 9 = 8. Lower fence = 9 − 1.5(8) = −3. Upper fence = 17 + 1.5(8) = 29. Since 42 > 29, 42 is an outlier. No values are below the lower fence. Part (b) — Boxplot construction (1 point) • (1 point) Correctly drawn modified boxplot with: box from Q₁ = 9 to Q₃ = 17; median line at 12; lower whisker from 9 to 5 (minimum); upper whisker from 17 to 28 (the largest non-outlier); and 42 plotted as an individual point beyond the upper whisker. Number line must be labeled. Part (c) — Display comparison (1 point) • (1 point) One valid reason for each. Example supporting histogram: A histogram would reveal the shape of the distribution more fully, including whether there are gaps or clusters in recovery times that the boxplot would obscure. Example supporting boxplot: The boxplot concisely displays the five-number summary and explicitly flags 42 as an outlier, making it easy to identify the extreme recovery time at a glance. Partial credit (0.5) for providing only one valid reason.

Question 5: CRITICAL THINKING

4 pts

Consider two datasets: Dataset X: 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 50, 50, 50, 50, 50, 50, 50, 50 Dataset Y: 10, 15, 18, 20, 22, 24, 26, 28, 30, 30, 30, 32, 34, 36, 38, 40, 42, 45, 48, 50 (a) Both datasets have the same minimum (10) and maximum (50), and similar medians. Explain why their boxplots could look similar despite the datasets having fundamentally different distributions. (b) Describe what feature of Dataset X's distribution would be completely hidden by a boxplot, and explain which alternative graphical display would reveal this feature. (c) Calculate the mean and standard deviation for Dataset X. Explain why these summary statistics, if displayed graphically (for example, by marking the mean and the interval from the mean minus one standard deviation to the mean plus one standard deviation on a dotplot of the data), would provide additional insight that the boxplot misses. (d) A student claims: "Since boxplots show the five-number summary, they always tell you everything important about a distribution." Evaluate this claim with reference to Datasets X and Y.

Show answer

Rubric (4 points): Part (a) — Explaining similar boxplots (1 point) • (1 point) The student explains that boxplots encode only five summary values (Min, Q₁, Median, Q₃, Max) and any datasets sharing similar values of these five numbers will produce similar-looking boxplots regardless of how the data are distributed between the quartiles. Dataset X is bimodal (clustered at 10 and 50 with nothing in between) while Dataset Y is approximately uniformly spread, but if the quartiles happen to be similar, the boxplots will not distinguish these cases. Part (b) — Hidden feature and alternative display (1 point) • (1 point) Dataset X is clearly bimodal — all values are either 10 or 50 with a large gap in between. A boxplot cannot reveal bimodality because it only shows quartile positions, not the distribution of values between quartiles. A histogram or dotplot would immediately show two distinct clusters of data with a gap from 11 to 49. Part (c) — Mean, SD, and graphical insight (1 point) • (1 point) For Dataset X: 12 values of 10 and 8 values of 50. Mean = (12 × 10 + 8 × 50)/20 = (120 + 400)/20 = 26. For the standard deviation: each 10 deviates by −16, each 50 deviates by 24. Σ(xᵢ − x̄)² = 12(256) + 8(576) = 3072 + 4608 = 7680. s = √(7680/19) ≈ √404.2 ≈ 20.1. If you mark x̄ = 26 and the interval x̄ ± s = (5.9, 46.1) on a dotplot, you would see that the mean falls in the gap where no data exist, and very few data points lie within one standard deviation — a dramatic visual signal that the distribution is not unimodal and that mean ± s does not describe a typical range. Part (d) — Evaluating the claim (1 point) • (1 point) The claim is incorrect. The five-number summary tells you about center (median), spread (IQR, range), and some information about skewness, but it cannot reveal modality (unimodal vs. bimodal), gaps, clusters, or the exact distribution of values between quartiles. Datasets X and Y demonstrate this limitation directly: they could produce similar boxplots but have fundamentally different distributional shapes. A complete description of a distribution requires assessing shape, center, spread, and unusual features — and boxplots alone cannot assess shape beyond crude skewness.

SUMMARY

Lesson Summary

Summary

Graphical representations of summary statistics transform numerical measures into visual displays that communicate distributional characteristics at a glance. The boxplot encodes the five-number summary (Min, Q₁, Median, Q₃, Max) and identifies outliers using the 1.5 × IQR rule, making it the most direct link between numerical summaries and visual form. Histograms reveal distributional shape — including modality and gaps — while dotplots preserve individual observations and can be annotated with mean and standard deviation markers.

When describing any distribution graphically, always address shape (symmetric, skewed left/right, unimodal/bimodal), center (mean or median), spread (IQR, standard deviation, or range), and unusual features (outliers, gaps, clusters) — and always interpret these in context. No single display reveals everything; the AP exam rewards students who can justify their choice of display and articulate what it shows and what it hides.