Table of Contents
Standard deviation measures how spread out data values are from the mean, calculated as the square root of variance. Population formula: σ = √(Σ(x − μ)² / N). Sample formula: s = √(Σ(x − x̄)² / (n − 1)). Small values cluster near the mean; large ones spread widely.
Standard deviation is a summary statistic showing how much individual values in a dataset typically deviate from the average. It turns a list of numbers into a single, comparable figure that describes consistency.
It is one of the most widely reported statistics in dissertations, lab reports and quantitative essays, because markers expect spread to be described alongside any average you quote.
A low standard deviation tells you most scores sit close to the mean. A high one tells you scores are scattered widely. Both readings are meaningful only when reported next to the mean itself, never alone.
You will see it used everywhere from comparing exam scores across cohorts to quality-control charts in manufacturing, describing how consistent a process or a group of results really is.
The concept sits inside a broader family of dispersion statistics. If you have not yet covered range, interquartile range and variance, our guide to measures of variability explains how these measures relate before you tackle standard deviation in depth.
Markers expect quantitative chapters to report spread, not just averages. Two datasets can share an identical mean yet behave completely differently — one tightly clustered, one wildly variable.
Standard deviation is what reveals that difference on paper. It also underpins other techniques you will meet later, including z-scores, hypothesis testing and confidence intervals.
Understanding it properly now makes those later chapters far easier to write and defend at viva. Examiners regularly ask students to justify why a spread measure was reported at all.
There are two versions of the formula, and choosing the right one matters for accuracy. Use the population formula only when your dataset is the entire group you care about.
Use the sample formula whenever your data is a sample drawn from a larger population — which covers almost every student survey, experiment or questionnaire dataset.
| Formula Type | Notation | When to Use |
|---|---|---|
| Population standard deviation | σ = √(Σ(x − μ)² / N) | Data covers the whole population |
| Sample standard deviation | s = √(Σ(x − x̄)² / (n − 1)) | Data is a sample used to estimate a population |
The only structural difference is the denominator: population divides by N, sample divides by n − 1. That adjustment, called Bessel’s correction, stops the sample formula underestimating true spread.
Every method, whether done by hand or in software, follows the same five-stage logic shown below. Working through it manually once makes SPSS or Excel output far easier to interpret later.
Variance and standard deviation measure the same underlying spread, but they are not interchangeable when writing up results. The table below sets out the differences students most often confuse.
| Feature | Variance | Standard Deviation |
|---|---|---|
| Definition | Average of squared deviations from the mean | Square root of variance |
| Units | Squared units (e.g. marks²) | Original units (e.g. marks) |
| Typical symbol | σ² or s² | σ or s |
| Ease of interpretation | Harder — units are not intuitive | Easier — directly comparable to the mean |
| Common use | Intermediate step in statistical tests (e.g. ANOVA) | Reported in results sections and abstracts |
In practice, most write-ups quote standard deviation. It stays in the same units as the original data, which makes it far easier for a reader to interpret at a glance.
Variance still matters, though. It is the figure calculators and software use internally, and it appears directly inside formulas for ANOVA and other significance tests.
For data that follow a roughly normal distribution, standard deviation follows a predictable pattern known as the empirical rule. This is what makes the statistic so useful for describing spread in one sentence.
Roughly 68% of values fall within one standard deviation of the mean, 95% within two, and 99.7% within three. This is sometimes called the 68-95-99.7 rule.
Values sitting more than two or three standard deviations away are usually flagged as outliers worth investigating, rather than automatically excluded from analysis.
These two terms are often confused but answer different questions. Standard deviation describes spread within one sample. Standard error describes how much a sample mean would vary across repeated samples.
Standard error is calculated by dividing standard deviation by the square root of the sample size, and it shrinks as sample size grows. This distinction matters directly when building a confidence interval around a mean.
As a rule of thumb: quote standard deviation to describe your own sample, and quote standard error when estimating how close your sample mean is to the true population mean.
Survey and questionnaire data often arrive as a frequency table rather than raw scores. The same logic still applies, but each squared deviation is multiplied by its frequency before summing.
Most students calculate this in SPSS or Excel rather than by hand once categories exceed five or six rows, since the manual arithmetic becomes repetitive and error-prone at scale.
Likert-scale questionnaires are a common example. Each response option (say, 1 to 5) has a frequency count, and standard deviation then shows how much opinion varied around the average rating.
Most UK marking criteria expect standard deviation reported in a consistent style, for example: M = 68.00, SD = 6.32. This single line communicates both average and spread without cluttering a results paragraph.
Report standard deviation in a table when comparing several groups, and in running text when discussing one variable. Match decimal places to whatever your department’s style guide requires, and stay consistent throughout.
Software removes the arithmetic but not the need to choose the right formula. In Excel, use =STDEV.P() for a population and =STDEV.S() for a sample.
In SPSS, request it under Analyze > Descriptive Statistics > Descriptives. In R, the sd() function calculates the sample version by default.
If your dissertation methodology involves SPSS output you are not confident interpreting, our SPSS data analysis help guide walks through reading descriptive tables before you write them up.
The same handful of errors appear in student work every year. Checking for these before submission catches most issues a marker would otherwise flag.
Get Help With Your Statistical Analysis
Standard deviation is one part of a wider statistics toolkit. Our guide to descriptive statistics places it alongside mean, median and mode.
For the full picture of spread measures and worked examples, browse our statistical analysis guide hub, where every technique in this article is covered in more depth.
Standard deviation is a number that shows how much values in a dataset typically differ from the average. A small standard deviation means data points sit close together; a large one means they are spread widely across a bigger range.
Find the mean of your dataset, subtract it from each value to get the deviations, then square every deviation so negatives cannot cancel positives. Average those squared deviations to find the variance, then take the square root of that variance to get the standard deviation.
For a population it is σ = √(Σ(x − μ)² / N). For a sample it is s = √(Σ(x − x̄)² / (n − 1)). The sample version divides by n − 1 instead of n to avoid underestimating true spread.
Variance is the average of squared deviations from the mean, expressed in squared units such as marks squared. Standard deviation is simply the square root of variance, expressed back in the original units, which is why most results sections report standard deviation rather than variance.
A high standard deviation means values are spread widely around the mean rather than clustered close to it. In a results section this signals more variability or inconsistency within the group being measured, and is often worth discussing alongside the mean when interpreting findings.
Use the population formula only if your data covers an entire population. Use the sample formula, which divides by n − 1, whenever your data is a sample used to estimate a wider population — this covers most student research.
You May Also Like
All work is written by human writers. 100% AI free, guaranteed.
100% money back guarantee if you find plagiarism in our work.
COMPANY DETAILS
