Standard Error Calculator

Two boxes, pre-filled with s = 15 and n = 25. Click Calculate. The standard error calculator form asks for exactly two numbers: the sample standard deviation s and the sample size n.

• s is the standard deviation of your one sample — how far individual values scatter. For proportion data, leave s empty and fill the p̂ box instead.
• n is the number of values averaged together in that sample, not the number of times you repeated anything.
• Four mean rows come back: SE = s/√n plus the 90%, 95%, and 99% margins of error. Fill the optional proportion box p̂ instead and the standard error calculator swaps in the proportion row.

The standard error calculator keeps results exact, so 15 and 25 return SE = 3 cleanly.

This page is more than a static standard error calculator — it is also an AI tutor you can ask questions:

• Ask in plain words. "Sample of 40 values with standard deviation 8 — what's the standard error?" The AI fills the form and runs this standard error calculator for you.
• Take a photo of your homework. The AI reads s and n from the problem and fills the standard error calculator boxes for you.
• Ask about the ideas, not just the answer. "Why does the standard error get smaller with bigger samples?" The AI explains step by step — every number still comes from the standard error calculator, so nothing is guessed.
• Keep asking follow-ups. Ask "why?" as often as you need — follow-up questions are welcome.
• Proportions and raw data. The form covers the mean; for a proportion or a raw data list, ask the AI on this standard error calculator page and it walks you there.

One measurement wobbles; an average steadies. That one sentence is the whole idea behind the standard error calculator. Take one sample of n values and compute its mean x̄. Take another sample. The two means differ — but far less than two single values would.

The picture shows why. The wide curve is the spread of single values, SD = 15. The tall curve is the spread of sample means from samples of n = 25; its standard deviation is the standard error, 15/√25 = 3. The standard error calculator measures that second curve with one division. It is "how far (on average) the sample mean will be from the population mean in repeated simple random samples."

Read the picture first: wide curve for values, tall curve for means. Now the symbols the standard error calculator uses:

$SE = \frac{s}{\sqrt{n}}$

Every number the standard error calculator returns comes from the SE row plus one multiplication.

A class of quiz scores has σ = 15 and you average n = 25 scores. Run the standard error calculator.

1. $SE = \frac{15}{\sqrt{25}} = \frac{15}{5} = 3$.
2. So the sample mean lands about 3 points from the true mean on average — the standard error calculator's whole message.
3. Two standard errors up: $90 + 2(3) = 96$ — exactly how a z question on the sample mean works.

Type 15 and 25 into the standard error calculator above and SE = 3 confirms instantly.

Scores on an exam: σ = 3, sample n = 36, sample mean x̄ = 68. Build the confidence interval — this is the standard error calculator's second job.

1. Standard error: $SE = 3/\sqrt{36} = 0.5$.
2. At 90% confidence, $MoE = 1.645(0.5) = 0.8225$.
3. Interval: $68 \pm 0.8225 = (67.18, 68.82)$. At 95%, $MoE = 1.96(0.5) = 0.98$, giving $(67.02, 68.98)$; at 99%, $MoE = 2.576(0.5) = 1.288$, giving $(66.71, 69.29)$.

Higher confidence buys a wider interval — the standard error calculator's three margin rows (90/95/99) show the trade side by side.

Quadruple the sample and let the standard error calculator show the payoff: n = 25 versus n = 100, both with s = 15.

1. At n = 25: $SE = 15/5 = 3$.
2. At n = 100: $SE = 15/10 = 1.5$.
3. Four times the data cut the standard error in half, because n sits under a square root.

The standard error calculator shows this instantly: change 25 to 100, click again, and the SE row halves. Precision is bought by √n, not by n — a fact worth feeling once.

Four slips the standard error calculator catches every day:

1. Confusing SD and SE. SD = 15 spreads single quiz scores; SE = 3 spreads sample means. Same data, different curves — the standard error calculator computes the second one.
2. Forgetting the square root. Dividing by n instead of √n gives 15/25 = 0.6 for the example above — off by fivefold. The shrinkage runs at √n speed.
3. Reading n as repeats. n is the number of values averaged together in one sample. Repeating an experiment k times does not enter the standard error calculator at all.
4. Quoting a margin with no confidence level. "Margin of error 0.98" means nothing alone; it is 0.98 at 95%. Always report the pair, exactly as the standard error calculator labels its rows.