Standard Error of the Mean Calculator
Compute the standard error SE = σ / √n — the uncertainty in a sample mean as an estimate of the population mean. Solve for SE, σ, or n; convert between SE and the standard deviation; check sample-size requirements; and see the sampling distribution live with a full step-by-step solution.
SE vs. SD — precision vs. spread
The standard deviation σ measures how spread out the individual observations are. The standard error SE measures how precisely the sample mean estimates the population mean. They are different quantities with different meanings: σ is a property of the data, SE is a property of the estimate. Confusing them is the single most common error in reporting. A large σ with a large n can still give a tiny SE — a noisy process can still produce a precise average if you measure it often enough.
SE shrinks as 1/√n — not as 1/n
The square-root law is the reason precision improves slowly. To halve the SE you need 4× the data; to reduce it by a factor of 10 you need 100× the data. This is why doubling a study's sample size yields only a 29 % improvement in precision. The Advanced chart shows this curve directly: SE drops steeply from n = 2 to n = 30 and then flattens — the gains become very expensive beyond a few hundred observations.
It defines the sampling distribution of x̄
The central limit theorem states that for reasonably large n, x̄ is approximately normally distributed about μ with standard deviation SE = σ/√n. This is the bridge that makes inference possible: it tells you exactly how much x̄ bounces around from one sample to the next, and hence how far off it might be from μ. Every confidence interval and every hypothesis test for a mean is built on this single fact.
SE depends on the population σ, not the sample s
Strictly, the standard error uses the population standard deviation σ. In practice σ is unknown and is replaced by the sample standard deviation s, giving the estimated standard error SE = s/√n. When n is small, s underestimates σ on average, so the t-distribution is used instead of z to widen the intervals and correct for that bias. As n grows, s converges to σ and the distinction disappears.
How to Use This Calculator
The standard error of the mean quantifies how precisely a sample mean estimates the population mean. Give the tool σ and n, and it computes SE — or solve for whichever of the three you do not know.
1. Load a scenario preset
Click any preset — IQ, SAT, quality control, polling, blood pressure, height — to load typical σ and n values in one click. Or type your own numbers directly.
2. Pick what to solve for
Choose SE to compute the standard error from σ and n. Choose σ to recover the standard deviation from a known SE and n. Choose n to find the sample size that would produce a given SE at a given σ.
3. Enter σ and n
σ is the standard deviation of the individual observations — the same units as the data. n is the number of observations. Both must be strictly positive, and n must be an integer ≥ 2.
4. Read the result and the diagram
The result card shows SE plus the 95% confidence-interval half-width E = 1.96·SE for reference. The diagram below overlays the wide population distribution with the much narrower sampling distribution of x̄, so the difference between σ and SE is visible at a glance.
5. Check the sampling distribution chart
The bell curve in the result column is the sampling distribution of x̄, with ±1, ±2, and ±3 SE bands shaded. About 68% of sample means fall within 1 SE of the true mean, 95% within 2 SE, 99.7% within 3 SE.
6. Switch to Advanced for depth
Advanced mode adds the SE-vs-n chart (showing the square-root law in action), a +1% sensitivity check on σ and n, and a metrics panel with the 90%, 95%, and 99% CI half-widths, the z and t critical values, the SE/σ ratio, and required n for several target SE values.
What This Tool Offers
Built around the fundamental sampling relation SE = σ/√n — the uncertainty of a mean estimate and the foundation of every confidence interval for μ.
Three-way solving
Solve for the standard error, the standard deviation, or the sample size. The calculator rearranges SE = σ/√n automatically for whichever variable you choose.
Population vs. sampling diagram
A live side-by-side overlay of the wide population distribution (σ) and the narrow sampling distribution of the mean (SE = σ/√n), with both curves drawn to scale so the precision gain is visible.
Sampling distribution chart
A Gaussian curve representing the distribution of x̄ with ±1 SE, ±2 SE, and ±3 SE bands shaded, so you can see directly how far x̄ typically falls from μ.
SE vs. n trade-off chart
Advanced mode plots SE = σ/√n against sample size n, revealing the square-root law: SE falls rapidly at small n and then flattens, making extra data progressively more expensive per unit of precision.
Full inference metrics
Beyond SE: the ratio SE/σ = 1/√n, the standard deviations, the 90%, 95% and 99% CI half-widths, the z and t critical values at several n, and required n for several target SE values.
Reference tables
A SE lookup table across σ and n, a required-n table for common SE targets, typical real-world σ and n values by application, and the full variable reference.
Reference Tables & Conversions
A SE lookup table across σ and n, a required-n table for target SE, typical σ values by application, and the variable reference.
Standard Error Lookup — SE = σ / √n
Standard error across a range of standard deviations σ and sample sizes n. Multiply any cell by σ/10 for a different σ.
| σ \ n | n = 4 | n = 9 | n = 25 | n = 100 | n = 400 | n = 1000 |
|---|
Required Sample Size for a Target SE
Computed from n = ⌈(σ/SE)²⌉. Assumes σ = 1 in the table — multiply SE by σ for other cases.
| Target SE (with σ = 1) | Required n | SE/σ | Precision class |
|---|
Typical Standard Deviations by Application
Representative σ values from real-world datasets, together with the SE at several sample sizes.
| Quantity | σ | Unit | SE at n = 10 | SE at n = 100 | SE at n = 1000 |
|---|
Variable Reference
| Symbol | Meaning | Formula | Notes |
|---|---|---|---|
| SE | Standard error of the mean | σ / √n | Uncertainty of x̄ as an estimate of μ |
| σ | Population standard deviation | — | Spread of individual observations |
| s | Sample standard deviation | √(Σ(xᵢ−x̄)²/(n−1)) | Used in practice when σ is unknown |
| n | Sample size | count | Positive integer ≥ 2 |
| x̄ | Sample mean | Σxᵢ / n | Point estimate of μ |
| μ | Population mean | — | Parameter being estimated |
| E | Margin of error | z · SE | Half-width of the confidence interval |
| 1/√n | SE-to-σ ratio | — | Precision factor; 0.5 at n = 4, 0.1 at n = 100 |
| √n | Square-root factor | — | How much σ exceeds SE |
What Does SE = σ / √n Measure?
The standard error of the mean quantifies how much the sample mean x̄ varies from one sample to the next. If you repeatedly drew samples of size n from the same population and computed the mean each time, those means would not be identical — they would scatter around the true population mean μ with a standard deviation equal to σ/√n. That scatter is the standard error. It is the fundamental quantity that connects a single sample to the population it came from, and it underlies every confidence interval, hypothesis test, and power calculation for a mean.
The Standard Error Formula
The standard error of the mean, when the population σ is known, is:
When σ is unknown and must be estimated from the sample as s, the estimated standard error is:
Rearranged to solve for the other two quantities:
n = (σ / SE)²
The SE is the natural unit of a confidence interval for the mean. In general:
And the SE-to-σ ratio is a simple, unit-free measure of the precision gained by sampling:
Standard Error Variables
| Symbol | Meaning | Notes |
|---|---|---|
| SE | Standard error of the mean | σ/√n; smaller means more precise |
| σ | Population standard deviation | Spread of the individual observations |
| s | Sample standard deviation | Estimated from the sample; used when σ is unknown |
| n | Sample size | Positive integer ≥ 2 |
| x̄ | Sample mean | Random variable with mean μ and SD SE |
| E | Margin of error | z or t critical value multiplied by SE |
How σ and n Affect the SE
The standard error depends on σ linearly and on n through its square root. Doubling σ doubles SE — a noisier process produces a less precise mean, as you would expect. But doubling n only reduces SE by a factor of √2 ≈ 1.414, or about 29%. This asymmetry is the defining feature of sampling: precision improves with the square root of the sample size, never linearly.
The practical consequences are stark. Starting from n = 10, going to n = 40 (four times the data) halves the SE. Going from n = 100 to n = 400 also halves the SE — but the second upgrade costs far more in absolute terms. This is why almost all survey research stops at around 1,000 to 2,000 respondents: the marginal precision gained by adding another thousand is tiny compared to the cost.
The ratio SE/σ = 1/√n makes this explicit. At n = 4 it is 0.5 — the mean is half as variable as the underlying data. At n = 100 it is 0.1 — one tenth as variable. At n = 10,000 it is 0.01 — one hundredth. Each order-of-magnitude improvement in precision requires a hundredfold increase in sample size.
Standard Error Calculation Examples
SE = 15 / √36 = 15 / 6 = 2.500 points
95% CI half-width E = 1.96 × 2.5 = 4.900 → mean is known to ± 4.9 IQ points
SE = 0.1 / √30 = 0.1 / 5.477 = 0.01826 mm
95% CI half-width E = 0.0358 mm → mean is known to ± 36 μm
σ = SE × √n = 2.0 × 5 = 10.0
n = (σ/SE)² = (15/1.0)² = 225 → n = 225 observations
Frequently Asked Questions
The standard deviation σ describes how spread out the individual observations are around the mean. The standard error SE describes how spread out the sample mean x̄ is around the population mean μ. They answer different questions and have different units of meaning. σ is a property of the data; SE is a property of the estimate. A dataset can have a large σ (very variable data) and a tiny SE (very precise mean) at the same time — you just need a large n. The rule is: σ is about the observations, SE is about the estimate.
Because variances add, not standard deviations. When you average n independent observations, the variance of the average is σ²/n. Taking the square root gives the standard deviation of the average, which is σ/√n. This is the fundamental square-root law of sampling and it appears everywhere in statistics. It means halving the standard error requires 4× the data, reducing it by 10× requires 100× the data, and reducing it by 100× requires 10,000× the data. There is no way around it — the only way to make SE small is a very large n.
Report SD when you are describing the spread of your data — how variable the observations are. Report SE when you are describing the precision of an estimated mean, or when you are illustrating a confidence interval. Both are legitimate; the mistake is reporting one while implying the other. A common problem is that SE is √n times smaller than SD, so error bars showing ± 1 SE look much tighter than error bars showing ± 1 SD — 5.5× tighter at n = 30. Some journals require SD, others allow either provided it is stated. Always label clearly.
A confidence interval is just the point estimate plus or minus a multiple of the standard error: CI = x̄ ± z·SE or x̄ ± t·SE. The multiple is the critical value from the standard normal or Student's t distribution, determined by the desired confidence level. So a 95% CI is x̄ ± 1.96·SE, a 90% CI is x̄ ± 1.645·SE, and a 99% CI is x̄ ± 2.576·SE. Everything about the interval's width comes from two sources: SE (which depends on σ and n) and the critical value (which depends on the confidence level and the distribution).
Only in degenerate cases. If σ = 0 every observation is identical, and the mean is known exactly — SE = 0. This almost never happens in real data. If n → ∞ the SE approaches 0, but never reaches it in a finite sample. Theoretically, if the population is finite and you sample the entire population (n = N), the standard error is exactly zero because the sample mean equals the population mean. For finite-population sampling without replacement, the correct formula includes the finite-population correction factor √((N−n)/(N−1)), which shrinks the SE.
There is no universal answer — it depends on σ (the variability of what you are measuring) and the precision you need. A useful starting point is to specify the SE you want and solve n = (σ/SE)². For example, to estimate a mean IQ to within ± 2 points (SE = 1.02 for a 95% CI half-width of 2), you need n = (15/1.02)² ≈ 217 subjects. For quality control on a manufacturing line with σ = 0.1 mm and a target SE of 0.02 mm, you need n = (0.1/0.02)² = 25 parts. The general pattern: precise estimates of noisy quantities require large samples.
When you sample without replacement from a finite population of size N, the observations are not quite independent, and the standard error is slightly smaller than σ/√n. The correction factor is √((N − n)/(N − 1)), and the corrected SE becomes SE = (σ/√n) · √((N − n)/(N − 1)). When n is small relative to N (say n/N < 0.05) the correction is negligible. When the sample is a large fraction of the population — say 30% or more — the correction matters and should be applied. This calculator assumes sampling with replacement (or an effectively infinite population).
Variables
| Symbol | Quantity | Unit |
|---|---|---|
| SE | Standard error of the mean | |
| σ | Population (or sample) standard deviation | |
| n | Sample size |
What it means
The standard error (SE) of the mean is the standard deviation of the sample means over many repeated samples from the same population. It quantifies the sampling variability of the mean. The formula shows that the SE decreases as the sample size increases. It is used to compute confidence intervals, to perform hypothesis tests (z‑tests, t‑tests), and to assess the reliability of sample estimates. Understanding the standard error is fundamental for inferential statistics and for interpreting sample results.
Worked example
Standard Error of the Mean – Two Examples
Real‑World| Parameter | Value |
|---|---|
| σ | 15 |
| n | 25 |
| Parameter | Value |
|---|---|
| σ | 10 |
| n | 100 |
Common mistakes
- Standard error of the mean (SE): The standard deviation of the sampling distribution of the sample mean.
- σ: Population standard deviation – if unknown, use s (sample SD) to estimate.
- n: Sample size – larger n gives smaller SE.
- Units: SE has the same units as the data.
- Interpretation: SE measures the precision of the sample mean as an estimate of μ.
Applications
The standard error of the mean (SE) is the standard deviation of the sampling distribution of the sample mean, measuring how much the sample mean is expected to vary from sample to sample. It is a key quantity in inferential statistics, used to construct confidence intervals and to perform hypothesis tests. Engineers use SE to assess the precision of experimental results, to compare groups, and to determine sample sizes. By calculating SE, professionals can quantify the reliability of the sample mean. A smaller SE indicates a more precise estimate. Understanding the standard error is essential for any work involving sampling, as it forms the basis for statistical inference and the interpretation of data.
- Sample size planning for experiments
- Confidence interval and hypothesis test construction
- Assessment of measurement precision
- Meta‑analysis – combining estimates from multiple studies
- Quality control – monitoring process mean stability
Related calculators · Probability Distributions
Recently updated formulas
Suggest an improvement
Found a mistake? Have a better example? Let us know.
You may send up to 3 suggestions per 24 hours.