AP Statistics: Quantitative Data Displays

Before you can analyze numerical data, you must first see it. Visualizing quantitative data—data that measures amounts or counts—is the critical first step in any statistical analysis. In AP Statistics, mastering the construction and interpretation of displays like histograms, dotplots, and stemplots is foundational; these tools transform a column of numbers into a story about shape, center, and spread, directly informing every hypothesis test and confidence interval you will build.

Histograms: The Shape of Distribution

A histogram is a bar graph that displays the frequency or relative frequency of quantitative data grouped into intervals, called bins or classes. It is the go-to display for visualizing the overall distribution of a dataset, especially with larger sample sizes (typically $n>30$).

To construct a histogram, you must first create a frequency table. Determine an appropriate number of bins (often between 5 and 20) and calculate the bin width: $(max-min)/(\text{number of bins})$. Each bin is a contiguous interval on the horizontal axis, and the height of the bar corresponds to the number (or proportion) of observations falling into that interval. Crucially, the bars in a histogram touch each other, emphasizing that the data is continuous.

The power of a histogram lies in its immediate revelation of a distribution's shape. You will describe shape using specific terms:

• Symmetric: The left and right sides of the distribution are approximately mirror images. A special case is the bell-shaped, normal distribution.
• Skewed: The data stretches out in one direction.
• Skewed right (positively skewed): The tail extends to the right. Think of personal income data—many people clustered at lower incomes, with a few very high incomes creating a long right tail.
• Skewed left (negatively skewed): The tail extends to the left. An example is exam scores where most students did very well, with only a few low scores.
• Unimodal: Having one clear peak.
• Bimodal/Multimodal: Having two or more clear peaks, which may suggest the data comes from two different groups (e.g., heights of adult men and women combined).

When describing a distribution in AP Statistics, you use the framework "CUSS": Center, Unusual features, Shape, Spread. From a histogram, you can make preliminary estimates of the center (the balancing point) and spread (the range of the data), and easily spot outliers—individual values that fall outside the overall pattern.

Dotplots: Simplicity for Smaller Sets

A dotplot is one of the simplest displays: a number line where each data point is represented by a dot stacked above its value. It is most effective for smaller datasets (typically $n<30$) where individual data points are meaningful and won't create an overwhelming pile.

Creating a dotplot is straightforward. Draw a horizontal axis covering the range of your data. For each observation, place a dot above the corresponding value on the axis. If a value repeats, stack the dots vertically.

The great advantage of a dotplot is that it preserves every single data point. You lose no detail to binning, which makes it perfect for identifying clusters, gaps, and precise outliers. You can see exactly how many times a specific value occurs and get an immediate sense of the mode (the most frequent value). For example, if you measured the reaction times of 15 students, a dotplot would let you see if two students had identical times and if one student's time was distinctly slower than the rest. While it shows shape, center, and spread, its primary strength is in detailing the granularity of the data.

Stemplots: A Detailed Look with Retention

A stemplot (or stem-and-leaf plot) is a hybrid display that organizes data while preserving the original values, much like a dotplot, but in a more structured, textual format. It is useful for small to moderately sized datasets and is excellent for finding the median and quartiles quickly.

To construct a stemplot, separate each data value into a "stem" (all but the last digit) and a "leaf" (the last digit). Write the stems in a vertical column. Then, for each data point, write its leaf in the row next to its stem. Leaves should be arranged in ascending order. For the data set {23, 25, 32, 34, 41}, the stem could be the tens digit and the leaf the ones digit:

2 | 3 5
3 | 2 4
4 | 1

This reads as the numbers 23, 25, 32, 34, and 41.

Like a dotplot, a stemplot shows the exact distribution of every data point, revealing shape, clusters, and gaps. Its ordered structure makes manual calculation of the five-number summary (minimum, $Q_1$, median, $Q_3$, maximum) and the interquartile range (IQR) straightforward. Stemplots can also be split or back-to-back to compare two distributions effectively. However, they become cumbersome with very large datasets or data that has many digits.

Selecting the Appropriate Display

Your choice of display is not arbitrary; it is a statistical decision based on the characteristics of your data and your analytical goals.

1. Based on Sample Size ($n$):

• Small $n$ ( < 30): Use a dotplot or stemplot. They preserve individual data points, which is valuable when you have few of them.
• Large $n$ ( ≥ 30): Use a histogram. Grouping data into bins simplifies the visualization and clarifies the overall shape, which is the priority with larger datasets.

1. Based on Analytical Goal:

• To see the overall shape and distribution pattern, use a histogram.
• To identify individual values, clusters, or exact gaps, use a dotplot or stemplot.
• To quickly find the median, quartiles, and IQR by hand, use an ordered stemplot.
• To compare two distributions, use back-to-back stemplots or adjacent dotplots/histograms.

The guiding principle is utility: which display will most clearly reveal the features you need to describe using CUSS?

Common Pitfalls

1. Misbinning in Histograms: Using too many bins creates a jagged, overly detailed plot that obscures the shape. Using too few bins oversimplifies and hides important features. Correction: Start with a rule of thumb like $\sqrt{n}$ for the number of bins, but always adjust to create a clear picture of the distribution. The bins must also be of equal width (unless you are using a density histogram, which is an advanced topic).

1. Confusing Skew Direction: Students often look at where the "hump" or mode is to determine skew. This is incorrect. Correction: Always look at the tail. If the longer tail points to the right (toward higher values), the distribution is skewed right. If the longer tail points to the left (toward lower values), it is skewed left.

1. Ignoring Context When Describing Spread: Stating "the spread is from 10 to 50" is incomplete. Correction: Always use a specific measure of spread in context. For example, "The range of exam scores was 40 points" or "The interquartile range (IQR) of reaction times was 0.3 seconds, meaning the middle 50% of times were within a 0.3-second interval."

1. Using the Wrong Display for the Sample Size: Creating a stemplot for 200 data points results in an unreadable list. Creating a histogram for 12 data points can artificially suggest a shape that isn't well-supported. Correction: Let the sample size guide your initial choice, as outlined in the previous section.

Summary

• Histograms use bins to show the shape of a distribution for larger datasets and are essential for describing patterns using the CUSS (Center, Unusual features, Shape, Spread) framework.
• Dotplots are ideal for smaller datasets, preserving every data point to reveal exact values, clusters, gaps, and outliers clearly.
• Stemplots organize data by place value, preserving individual values while providing an ordered structure that facilitates the calculation of the median, quartiles, and interquartile range.
• Your choice of display is a key analytical decision: use dotplots or stemplots for small $n$ to see details, and histograms for large $n$ to see the overall shape.
• Always describe a distribution in context, carefully identifying its shape (symmetric, skewed), center (mean or median), spread (range, IQR, or standard deviation), and any unusual features like outliers or gaps.