Describing Distributions

When statisticians collect quantitative data — measurements like heights, test scores, or reaction times — the raw numbers alone tell very little. To make sense of a dataset, we describe its distribution: the pattern of values and how often they occur. A distribution answers the fundamental question: 'What does this data look like, and what is typical?' The four key features every statistician examines are shape, center, spread, and unusual features (such as outliers or gaps). These four features work together. Knowing only the center (say, the average test score is 74) ignores whether scores are tightly clustered or wildly scattered, and whether the distribution is symmetric or skewed. A complete description requires all four elements. This framework is sometimes summarized by the acronym SOCS: Shape, Outliers, Center, Spread — a checklist statisticians use whenever they encounter a new dataset. Describing distributions is not merely a mechanical exercise. The choice of which measure of center or spread to report depends on the shape of the distribution. A strongly skewed distribution or one with outliers calls for the median and interquartile range, while a symmetric, bell-shaped distribution is well summarized by the mean and standard deviation. Understanding why these choices matter is a central goal of data exploration.

Key takeaways

Shape
Center
Spread
Unusual Features

Shape

Shape describes the overall pattern of a distribution as seen in a histogram or dot plot. A symmetric distribution is a mirror image on both sides of its center — the classic bell curve is symmetric. A right-skewed (positively skewed) distribution has a long tail stretching toward higher values; most data cluster on the left with a few unusually large values pulling the tail right. A left-skewed (negatively skewed) distribution has a long tail toward lower values. Distributions may also be uniform (all values roughly equally frequent), bimodal (two distinct peaks), or multimodal (more than two peaks). Identifying the number of modes (peaks) is part of describing shape.

Center

Center is a single value that represents a 'typical' observation. The two main measures are the mean and the median. The mean is the arithmetic average: add all values and divide by the number of values, giving x̄ = (Σxᵢ)/n. The median is the middle value when data are ordered; it splits the distribution so that 50% of values fall on each side. For symmetric distributions, mean ≈ median. For right-skewed distributions, the mean is pulled toward the long tail and exceeds the median. For left-skewed distributions, the mean is pulled left and falls below the median. The direction the tail pulls the mean away from the median is a useful diagnostic.

Spread

Spread (also called variability or dispersion) measures how much values differ from one another. Range = maximum − minimum, but it is sensitive to a single extreme value. The interquartile range (IQR) = Q3 − Q1, where Q1 is the 25th percentile and Q3 is the 75th percentile; IQR captures the spread of the middle 50% of data and is resistant to outliers. Standard deviation (s) measures the typical distance of observations from the mean and is best used with symmetric distributions. Variance is the square of the standard deviation. A large spread means data are highly variable; a small spread means values are tightly packed around the center.

Unusual Features

After describing shape, center, and spread, look for unusual features. Outliers are individual values that fall far from the bulk of the data. A common rule identifies a potential outlier as any value below Q1 − 1.5 × IQR or above Q3 + 1.5 × IQR. Gaps are regions with no data between clusters of values, sometimes signaling that two separate groups were combined. Clusters suggest subgroups in the data. Always investigate unusual features: they may reveal data-entry errors, genuinely interesting phenomena, or the need to split the dataset into subgroups before further analysis.

The Key Equation

\bar{x} = \frac{1}{n}\sum_{i=1}^{n} x_i, \quad s = \sqrt{\frac{\sum_{i=1}^{n}(x_i - \bar{x})^2}{n-1}}, \quad \text{IQR} = Q_3 - Q_1

\bar{x} — Sample mean (same as data): Arithmetic average; balancing point of the distribution
n — Sample size (dimensionless): Total number of observations in the dataset
x_i — Individual observation (same as data): The i-th data value in the dataset
s — Sample standard deviation (same as data): Typical distance of an observation from the mean; uses n−1 (Bessel's correction)
Q_1, Q_3 — First and third quartiles (same as data): 25th and 75th percentiles; divide the ordered data into quarters
\text{IQR} — Interquartile range (same as data): Spread of the middle 50% of the data; resistant to outliers

Worked Example

Problem. A student records the daily high temperatures (°F) for 11 days in April: 52, 55, 57, 58, 60, 62, 63, 65, 67, 70, 91. Describe the distribution using SOCS.

\text{Median} = 62^\circ\text{F},\quad \text{IQR} = 10^\circ\text{F},\quad \text{Outlier: } 91 > Q_3 + 1.5\times\text{IQR} = 82^\circ\text{F}

Answer. Shape: roughly symmetric with one right outlier; Center: median = 62°F (mean ≈ 63.6°F); Spread: IQR = 10°F, range = 39°F; Outlier: 91°F (exceeds upper fence of 82°F).

Figure

Diagram

Origins · 1 minJohn Tukey and how this was discovered

John Tukey

1915–2000 · American

John Tukey revolutionized data analysis by developing exploratory data analysis (EDA) — the practice of examining data graphically and numerically before formal inference. He invented the box-and-whisker plot (boxplot), which visually displays the five-number summary and uses the 1.5 × IQR rule to flag outliers. His 1977 book 'Exploratory Data Analysis' established the framework that statisticians still use today for describing distributions. Tukey coined the word 'software' in 1958 in a statistical journal, predating its widespread use in computing.

Histograms and Dot Plots — These graphical displays are the primary tools used to visualize shape, spread, and outliers when describing a distribution.
Measures of Center: Mean, Median, and Mode — Center is one of the four SOCS features; understanding when to use the mean versus median requires knowledge of skewness and outliers.
Measures of Spread: Range, IQR, and Standard Deviation — Spread quantifies how variable a distribution is and which measure to use depends directly on the shape of the distribution.
Boxplots and the Five-Number Summary — A boxplot is a compact visual summary of center, spread, and outliers based on Q1, median, Q3, and the 1.5×IQR rule.
The Normal Distribution — The symmetric bell-shaped normal distribution is the most studied distribution shape, and many statistical methods assume data are approximately normal.