Standard Error vs Standard Deviation

Confused by two numbers that sound almost the same? This guide explains the difference with one worked example, a side-by-side table and clear rules for error bars — so you always report the right one.

Standard deviation describes how spread out individual values are; standard error describes how precisely the sample mean estimates the population mean. For the sample 10, 12, 14, 16, 18, the standard deviation is 3.162 and the standard error is 1.414 — the same data, two different questions.

Standard Error vs Standard Deviation: The One-Line Difference

Standard deviation (SD) tells you how spread out the individual values in a data set are. Standard error (SE) tells you how uncertain the sample mean is as an estimate of the population mean. Same data, different questions: SD describes the data you collected, while SE describes the estimate you made from it.

Standard deviation answers “How spread out are the values?” Standard error answers “How sure are we about the mean?” Keep that pair of questions in mind and you will rarely mix the two up again.

What Is Standard Deviation?

The standard deviation is the typical distance between a value and the mean of the data set. A small standard deviation means the values cluster tightly around the mean; a large one means they are widely spread. It is a descriptive statistic: it summarises the data in front of you and says nothing, by itself, about any wider population.

Sample standard deviation: s = √[ Σ(x − x̄)2 / (n − 1) ]

The denominator (n − 1) instead of n is Bessel’s correction: it makes the sample variance an unbiased estimate of the population variance. The population standard deviation, σ = √[ Σ(x − μ)2 / N ], divides by N instead. The two versions differ only by that denominator — see our guide to population vs sample standard deviation and the step-by-step how to calculate standard deviation worked example.

What Is Standard Error?

The standard error is the standard deviation of the sample means. If you repeated the same study many times, the means of all those samples would vary from one sample to the next, and the standard deviation of those means is the SE. It measures the precision of your sample mean as an estimate of the population mean, and it depends on both the spread of the data and the size of the sample:

Standard error: SE = s / √n

Because the SE divides by √n, larger samples give smaller standard errors and more precise estimates of the mean. The data can be just as spread out, yet the mean becomes better pinned down. Try it yourself with the Standard Error Calculator — and see why the sampling distribution of the mean narrows as n grows in our guide to the central limit theorem.

Standard Error vs Standard Deviation: Comparison Table

The table sums up every difference that matters in practice. Read it row by row and the confusion between the two statistics starts to disappear.

 Standard deviation (SD)Standard error (SE)
What it measuresSpread of individual values around their meanUncertainty of the sample mean as an estimate of the population mean
Descriptive or inferential?Descriptive: it describes the sample you collectedInferential: it supports conclusions about the wider population
Formulas = √[ Σ(x − x̄)2 / (n − 1) ]SE = s / √n
Affected by sample size?Barely: it estimates the population spread, so it settles near a stable value as n growsYes, directly: it shrinks in proportion to 1 / √n, so bigger samples give smaller standard errors
Size compared with the otherLarger whenever n > 1Always smaller than the SD whenever n > 1
UnitsSame units as the dataSame units as the data (and as the mean)
Used forDescribing and summarising how variable a data set isConfidence intervals, hypothesis tests and margins of error
In a chartSD bars show the spread of the dataSE bars show the precision of the mean

Worked Example: Standard Deviation and Standard Error from the Same Data

To see the two numbers side by side, use one small data set for both. The data set is 10, 12, 14, 16, 18.

  1. Mean. (10 + 12 + 14 + 16 + 18) ÷ 5 = 70 ÷ 5 = 14.
  2. Deviations from the mean. 10 − 14 = −4; 12 − 14 = −2; 14 − 14 = 0; 16 − 14 = 2; 18 − 14 = 4.
  3. Squared deviations. 16, 4, 0, 4, 16. Their sum is 16 + 4 + 0 + 4 + 16 = 40.
  4. Sample variance and sample standard deviation. Sample variance = 40 ÷ (5 − 1) = 40 ÷ 4 = 10. Sample SD s = √10 = 3.162.
  5. Standard error. SE = 3.162 ÷ √5 = 3.162 ÷ 2.236 = 1.414.
  6. Population standard deviation, for comparison. σ = √(40 ÷ 5) = √8 = 2.828. The standard error from it is 2.828 ÷ 2.236 = 1.265 — smaller than 1.414, because dividing by 5 instead of 4 gives a smaller spread.
  7. 95% confidence interval for the mean. 14 ± 1.96 × 1.414 = 14 ± 2.77 = [11.23, 16.77].
QuantityValue
Mean14
Sample standard deviation (s)3.162
Standard error (SE)1.414
Population standard deviation (σ)2.828
Standard error from the population SD1.265
95% confidence interval for the mean[11.23, 16.77]

Notice the pattern: the SD (3.162) is more than twice the SE (1.414) even though both describe this one small sample. That gap grows as the sample grows. Check your own numbers with the Standard Deviation Calculator and try the SE calculator on the same values.

The Relationship Between SD and SE

The two statistics are linked by one equation, so you can convert between them whenever you know the sample size:

SD = SE × √n   and   SE = SD / √n

Check it with the worked example: 1.414 × √5 = 1.414 × 2.236 = 3.162, which is exactly the sample standard deviation.

Because the SE divides by √n, the gain from a bigger sample is slower than most people expect. This is the quadrupling rule: to halve the standard error you must multiply the sample size by four. Doubling the sample only cuts the SE by about 29%, because 1 / √2 = 0.707. With the worked example’s spread (s = 3.162):

Sample size (n)Standard error (SE)Change
51.414—
200.707Halved (4 × the sample)
800.354Halved again
3200.177Halved again

Meanwhile the standard deviation itself barely moves as n grows: it is an estimate of the population spread, so it settles near the true value rather than shrinking. The same 1 / √n logic applies to proportions — the standard error of a sample proportion is √[ p(1 − p) / n ] — which is why the sampling distribution of the sample proportion calculator shows the same narrowing pattern. The SE is also the building block of the margin of error; see margin of error vs confidence interval for how the two are connected.

When to Use Standard Deviation vs Standard Error

Use the statistic that answers the question you are actually asking.

Use the standard deviation when you want to:

  • Describe how spread out the values in a data set are — for example, “scores were 72 ± 9 (mean ± SD)”.
  • Summarise the results of a study or compare the variability of two groups or two measurement methods.
  • Feed other calculations that need the spread itself, such as a z-score or the coefficient of variation.

Use the standard error when you want to:

  • Estimate a population mean from a sample mean.
  • Build a confidence interval for the mean: mean ± a multiple of the SE (1.96 for a large sample, a t value for a small one).
  • Run or report a hypothesis test about the mean.
  • Report the margin of error of a survey estimate.

A handy rule of thumb: SD describes the sample; SE describes the estimate. If your sentence is about the data, use the SD. If it is about how well you know the population mean, use the SE.

Error Bars Explained: SD Bars vs SE Bars vs CI Bars

Error bars are the most misunderstood part of a statistical graph, and the bar you choose changes what the reader concludes. All three types are drawn around the same mean, but they mean very different things:

Bar typeWhat it showsLength, for the worked example (mean 14)What the reader learns
SD barsThe spread of the individual values14 ± 3.162 = [10.84, 17.16]How variable the data are
SE barsThe precision of the mean14 ± 1.414 = [12.59, 15.41]How precisely the mean is known
95% CI barsThe interval likely to contain the population mean14 ± 1.96 × 1.414 = [11.23, 16.77]Where the true mean probably lies

Why SE bars look deceptively small. SE bars are about √n times shorter than SD bars: with n = 5 they are about 2.24 times shorter, and with n = 100 they are 10 times shorter. A chart with SE bars therefore looks far more precise than the underlying data really are. Always label the bar type in the figure caption or legend, and state n alongside it.

Which bars should you draw? Show the spread of the data with SD bars (or a dot plot or box plot). Show how precisely the mean is estimated with SE or CI bars. And if you are comparing two means, prefer CI bars: overlapping SE bars do not prove there is no difference — that question needs a proper test, not a glance at the bars.

The confidence interval itself is built from the standard error: the NIST/SEMATECH e-Handbook gives the confidence limits for the mean as the sample mean plus and minus a multiple of the standard error.

The Common Mistake: Reporting SE Instead of SD

The most common error in published research is to report the standard error where the standard deviation belongs — describing the variability of the data with a number that actually describes the precision of the mean. Because the SE is always smaller (for n > 1), the data look less variable than they really are and the results look more impressive than they are.

Published analyses of the research literature have found this mix-up in a large share of papers — in some journals, in up to 64% of the articles examined — which is why reviewers and careful readers always check what the “±” actually is. When you read a paper, look for the words: “mean ± SD” describes the data, while “mean ± SE” (or “mean (SE)”) describes the estimate. When you write one, label the error bars, name the statistic after every “±”, and state n next to every mean.

Is Standard Error Always Smaller Than Standard Deviation?

Yes — whenever the sample has more than one value. Because SE = s / √n and √n is greater than 1 for every n > 1, dividing the SD by √n always gives a smaller number. In the worked example, the SD is 3.162 and the SE is 1.414. The two are exactly equal only when n = 1, because √1 = 1 — and with a single value there is no spread to describe anyway. The SE can never be larger than the SD: if you ever see a reported SE bigger than its SD, one of the two numbers is wrong, or the “±” was mislabelled.

Should I Use Standard Deviation or Standard Error in My Graph?

It depends on what you want the reader to take away. Use SD bars when the point of the graph is the spread of the data — how much individual values vary. Use SE bars (or, better, 95% CI bars) when the point is how precisely the mean has been estimated. Never draw SE bars and let the reader assume they show the data’s spread: for the worked example, SE bars span [12.59, 15.41] while the values themselves range from 10 to 18, so the SE bars hide most of the variation. Whatever you choose, label the bars clearly in the figure caption and give n.

What Is the Difference Between Standard Error and Standard Deviation in Simple Terms?

Think of a class of students. The standard deviation is how much the students’ test scores differ from each other — the spread of the class. The standard error is how certain you are about the class average — the precision of the mean. A class can have widely spread scores (a large SD) yet a very precisely known average (a small SE) if the class is large, because averages settle down as you add more students. One number describes the people; the other describes your estimate.

Frequently Asked Questions

What is the difference between standard error and standard deviation?

The standard deviation (SD) measures how spread out the individual values in a data set are around their mean. The standard error (SE) measures how precisely the sample mean estimates the population mean. In the worked example, the SD is 3.162 and the SE is 1.414 — the same data, two different questions.

Is standard error always smaller than standard deviation?

Yes, for any sample with more than one value (n > 1), because SE = s / √n and √n is greater than 1. The two are equal only when n = 1, and the SE can never be larger than the SD.

Should I use standard deviation or standard error in my graph?

Use SD bars to show how the data vary, and SE or 95% CI bars to show how precisely the mean is estimated. Always label the bars and give the sample size n, because SE bars are about √n times shorter than SD bars and can make the data look more precise than they are.

How do you calculate standard error from standard deviation?

Divide the standard deviation by the square root of the sample size: SE = s / √n. For the worked example, SE = 3.162 / √5 = 3.162 / 2.236 = 1.414.

What is the standard error if the standard deviation is 3.162 and n = 5?

The standard error is 1.414. Divide 3.162 by √5 (2.236): 3.162 / 2.236 = 1.414.

Why does standard error get smaller when the sample gets bigger?

Because SE = s / √n: the sample size is in the denominator under a square root. Quadrupling the sample halves the SE — with s = 3.162, the SE falls from 1.414 at n = 5 to 0.707 at n = 20 and 0.354 at n = 80. The standard deviation itself stays roughly the same.

When should I report the standard error instead of the standard deviation?

Report the SE when you are estimating a population mean, building a confidence interval, running a hypothesis test, or giving a margin of error. Report the SD when you are describing how variable the data themselves are.

What is the 95% confidence interval for a mean of 14 with a standard error of 1.414?

It is [11.23, 16.77]. Multiply the standard error by 1.96 (1.96 × 1.414 = 2.77) and add and subtract it from the mean: 14 ± 2.77 = [11.23, 16.77].

Check Tools

Try the numbers from this guide on your own data:

Standard Deviation Calculator

Find the SD of any data set, with the step-by-step working shown.

Try the Standard Deviation Calculator →

Standard Error Calculator

Work out the SE from your data, or from a known SD and sample size.

Try the Standard Error Calculator →

Confidence Interval from Mean and Standard Deviation Calculator

Build the 95% confidence interval for a mean, exactly as in the worked example above.

Try the confidence interval calculator →

You can find every calculator on the site in one place: Browse all our statistical tools →