Direct answer
Standard error describes how an estimator varies across repeated samples; for an independent sample mean it is σ/√n when population spread is known, while the central limit theorem can justify an approximate normal sampling distribution under suitable conditions.
Visual explanation
Samples create a distribution of estimates
What this calculation tells you
Sampling variability is the change in a statistic caused by observing one random sample rather than another. A sampling distribution describes that repeated-sample behavior.
Define the estimator, sampling unit, dependence structure, and target population before using a standard-error formula. Match the formula to the design rather than inserting n mechanically.
Where it is used
Survey research
Distinguish respondent variability from uncertainty in a sample estimate.
Experiments
Understand precision gains from repeated independent units.
Quality
Separate process spread from sampling uncertainty in a mean.
Education
Visualize the repeated-sampling definition of inference.
Common situations
- Explaining why two samples give different means.
- Calculating a simple mean standard error.
- Evaluating what larger n changes.
- Recognizing when dependence invalidates √n reasoning.
Start with the statistical question
Define the estimator, sampling unit, dependence structure, and target population before using a standard-error formula. Match the formula to the design rather than inserting n mechanically.
Repeated samples are drawn from one population; their means accumulate into a second distribution whose width narrows as n increases, while the raw population shape remains unchanged.
Worked example
For independent observations with population SD 15, a sample mean based on n=25 has SE 15/√25=3. Increasing to n=100 halves the SE to 1.5 rather than dividing it by four.
Assumptions that carry the result
The elementary formula assumes independent observations or an appropriate equivalent design. Clustered, serially correlated, weighted, finite-population, and complex survey estimates need design-aware variance methods.
Interpret the result without overreaching
A small standard error does not remove measurement error, selection bias, confounding, model misspecification, or a poorly defined target population.
- Using standard deviation and standard error interchangeably.
- Claiming the CLT makes observations normal.
- Treating a large sample as protection against bias.
Practical questions
Frequently asked questions
Does standard error measure data spread?
It measures sampling spread of a statistic; standard deviation measures spread among observations.
Does the CLT require normal raw data?
No, but its approximation quality depends on sample size, dependence, tail behavior, and the statistic.
Can a precise estimate be biased?
Yes. Sampling variation and systematic bias are distinct.
Further reading
Authoritative sources
Use these primary and professional resources to check definitions, conventions, or requirements that may extend beyond this guide.