Part of the Maths & Statistics suite · 235 calculators

Margin of Error Calculator

The ± figure quoted after every poll — what it means, how sample size drives it, and why the difference between two candidates has a bigger one.

For a proportion, the margin of error is z × √(p(1 − p)/n).

50% is the most conservative
%
Results update as you type
Results
Margin of error
3.099%
Interval — lower
Interval — upper
Margin on a difference between two shares
Z-value used
Sample needed to halve this margin
Sample needed for ±3 points
Reviewed September 2026. Pure mathematics: the result does not depend on where you are. Terminology follows the Australian Curriculum (maths, brackets, decimal point).
No account required · Google Analytics off unless allowedCalculator arithmetic runs in your browserResults update as you type
All calculations run 100% in your browser. The calculator code does not submit your figures to GlobalCalc to obtain a result.
About margin of error

How the margin of error calculator works

For a proportion, the margin of error is z × √(p(1 − p)/n). At 95% confidence, p = 50% and n = 1,000, that is about ±3.1 points — which is why almost every national poll quotes ±3.

The number applies to a single proportion. The margin on the *difference* between two candidates is roughly twice as large, which is why "within the margin of error" is misused constantly.

Formula: MoE = z √(p(1 − p) / n)

Worked examples

InputsMargin of errorNote
A poll of 1,0003.099%±3.1 points
A poll of 4004.8999%±4.9 points
A poll of 4,0001.5495%±1.55 points — four times the cost, half the margin

Frequently asked questions

Why is the margin of error almost always ±3?

Because almost every national poll samples about 1,000 people, and 1,000 at 95% confidence gives ±3.1 points.

Does the population size matter?

Barely. Polling 1,000 people gives the same precision in a country of 5 million as in one of 300 million — a fact that surprises almost everyone.

What does "within the margin of error" mean?

Usually it is misused. The quoted margin applies to one share; the margin on the *gap* between two candidates is about twice as large, so a 4-point lead in a ±3 poll is not outside it.

Does it cover every kind of error?

No — only random sampling error. Non-response bias, question wording and bad weighting are not in the number, and they are usually the bigger problem.

How big a sample do I need for ±1 point?

About 9,600 at 95% confidence. Precision is expensive: the cost rises with the square of the precision.

Where these figures come from

Last checked: September 2026. Formulas are fixed by mathematics and do not change with tax years or regulations.