Sampling variability

Standard Error of the Mean Calculator

Estimate the standard deviation of a sample mean from the population standard deviation and sample size.

Sampling inputs

Enter deviation and sample size

SEM

Population standard deviation

Sample size

Sampling summary

Calculated standard error

The standard error will appear here

Sample size must be a whole number of at least one.

Several sample distributions around a population mean becoming narrower with a larger sample size — For the same population spread, increasing sample size reduces the standard error.

What is Standard Deviation?

This calculator finds the standard deviation of the sampling distribution of a mean, commonly called the standard error of the mean when the population standard deviation is known. Enter the population standard deviation and the independent sample size. The result describes how much sample means would vary across repeated samples under the model.

It does not calculate the standard deviation of raw observations and it does not find a sample mean. It also does not turn poor data into reliable evidence: sampling design, dependence, bias, outliers and the suitability of the population model remain important.

How to Use This Calculator

Enter a nonnegative population standard deviation, σ, in the same unit as the measured variable.
Enter the positive whole-number sample size, n.
Select Calculate.
Read σ divided by the square root of n. The display rounds to four decimal places.
Use the unrounded value when building a confidence interval or doing further arithmetic.

The formula assumes the observations contributing to the mean are independent and have a common finite variance. If observations are clustered, paired, serially correlated or collected with unequal weights, an effective sample size or a method suited to that design may be needed.

Formula Used

Standard deviation of the sample mean = σ_x̄ = σ ÷ √n

Variance scales as σ²/n, so taking the square root gives σ/√n. If the population standard deviation is unknown and replaced by a sample standard deviation, the calculated number is usually called an estimated standard error. Small-sample inference commonly uses a t distribution rather than treating that estimate as a known σ.

The square-root relationship means quadrupling n halves the standard error; merely doubling n reduces it by a factor of √2.

Variables Explained

Symbol or term	Meaning
σ	Population standard deviation entered in the measured variable’s unit.
n	Positive whole number of independent observations in each sample.
x̄	Sample mean, whose repeated-sample variation is being described.
σ_x̄	Standard deviation of the sampling distribution of x̄.
Standard error	A standard deviation for a statistic’s sampling distribution.
√n	Square-root sample-size factor that reduces variation in the mean.

Worked Example

Suppose a process has a known population standard deviation of 12 units, and each sample contains 36 independent observations. The square root of 36 is 6. Divide 12 by 6:

σ_x̄ = 12 ÷ √36 = 12 ÷ 6 = 2

The calculator displays 2.0000. This says that repeated sample means have a standard deviation of 2 units under the assumptions. It does not say every sample mean lies within 2 of the population mean. Probabilities or confidence intervals require a distribution and a chosen confidence level.

Why Use This Calculator?

A smaller standard error means the sample mean is more stable across repeated samples, not that the observations themselves have become less variable. The result keeps the original measurement unit: if σ is 12 grams, the standard error is 2 grams.

Increasing sample size gives diminishing returns because precision improves with √n. To cut standard error in half while σ stays fixed, four times as many independent observations are needed. A tiny standard error can still accompany selection bias or systematic measurement error. For non-independent data, using the raw row count as n often understates uncertainty. When σ is estimated, distinguish descriptive calculation from the inferential procedure chosen for the study.

The formula is especially useful for planning comparisons because it makes the precision-versus-sample-size tradeoff visible. If σ = 12 and n = 36, increasing n to 144 reduces the standard error from 2 to 1. Reaching 0.5 would require n = 576 under the same assumptions. That arithmetic does not decide whether the extra observations are worth their cost, and it assumes new observations add independent information. Repeated measurements on the same unit may improve measurement precision, yet they are not automatically equivalent to observations from new independent units.

Common Mistakes

Entering variance instead of standard deviation.
Using zero, a decimal or a negative number for sample size.
Dividing by n rather than √n.
Calling the answer the spread of individual observations.
Assuming a large n removes sampling or measurement bias.
Counting correlated observations as fully independent.
Rounding the standard error before using it in later calculations.

Frequently Asked Questions

Is this the same as the standard deviation of the data?

No. It describes variation among repeated sample means, which is smaller by the √n factor.

What if the population standard deviation is unknown?

A sample standard deviation can estimate the standard error, but inference then needs methods that account for that estimation.

Why must sample size be an integer?

It counts observations, so partial observations are not meaningful here.

Does a smaller standard error prove the estimate is unbiased?

No. It describes repeated-sample spread and does not detect systematic bias.

What happens if I quadruple sample size?

The standard error is halved when population standard deviation and independence remain unchanged.

What unit does the answer use?

The same unit as the population standard deviation and measured variable.

Can I use the row count for clustered data?

Not automatically. Dependence can reduce effective information, so use a method appropriate to the sampling design.