Measures of Variability

Measures of variability, or dispersion, quantify how spread out a set of observations is around its centre. Two datasets can share the same mean yet differ greatly in how tightly their values cluster, and measures such as the range, variance, standard deviation, and interquartile range capture that difference.

Definition

A measure of variability quantifies the spread of observations around a central value: the range is the difference between the largest and smallest values, the variance is the mean squared deviation from the mean, the standard deviation is its square root in the original units, and the interquartile range is the spread of the middle half of the ordered data.

Scope

This entry covers the principal measures of dispersion — range, variance, standard deviation, and interquartile range — and how each is computed and interpreted. It distinguishes the standard deviation from the standard error and is a methodological reference, not clinical guidance.

Core questions

• How widely do the observations spread around their centre?
• Which dispersion measure pairs appropriately with the chosen measure of location?
• How does the standard deviation differ from the standard error?

Key concepts

• Range
• Variance
• Standard deviation
• Interquartile range
• Coefficient of variation
• Standard deviation versus standard error
• Pairing dispersion with central tendency

Mechanisms

The range, the gap between the extremes, is simple but unstable because it depends on only two values and grows with sample size. The variance averages the squared deviations of observations from the mean, and the standard deviation returns that quantity to the original measurement units, making it the natural companion of the mean for approximately symmetric data. The interquartile range, spanning the 25th to 75th percentiles, describes the middle half of the data and is robust to outliers, making it the companion of the median for skewed distributions. A recurring source of confusion is the difference between the standard deviation, which describes the spread of individual observations, and the standard error, which describes the precision of an estimate such as the mean and shrinks as the sample grows.

Clinical relevance

Dispersion measures tell readers how variable a measurement or outcome is, which matters for judging consistency, reference ranges, and the precision of reported estimates. This entry describes how variability is summarised for appraisal and is not a basis for individual diagnostic or treatment decisions.

Epidemiology

Reporting variability alongside central tendency is a basic expectation in health research, and the standard-deviation-versus-standard-error distinction is a common reporting error: confusing them can make estimates look more or less precise than they are. The interquartile range is preferred when data are skewed.

History

The variance and standard deviation were formalised in the late nineteenth and early twentieth centuries, with the term standard deviation introduced by Karl Pearson and the analytic framework of variance developed by Ronald Fisher. The robust, quantile-based interquartile range gained prominence with the rise of exploratory data analysis and the box plot in the twentieth century.

Debates

Standard deviation or standard error in reporting?

Authors frequently report the standard error in place of the standard deviation because it is numerically smaller, which can mislead readers about the variability of the underlying observations; methodological guidance stresses reporting the standard deviation to describe spread and reserving the standard error for the precision of estimates.

Key figures

• Douglas G. Altman
• J. Martin Bland
• S. Manikandan

Seminal works

• manikandan-2011-dispersion
• altman-bland-2005

Frequently asked questions

What is the difference between the standard deviation and the standard error?

The standard deviation describes how much individual observations vary around the mean, while the standard error describes how precisely the mean itself is estimated. The standard error decreases as the sample size increases; the standard deviation does not.

When should the interquartile range be used instead of the standard deviation?

When the data are skewed or contain outliers, the interquartile range describes spread more faithfully because, like the median, it is unaffected by extreme values.