Sample Size Calculator
Choose how precisely you need to estimate a percentage or a mean. The calculator returns the required completed responses, applies a finite population correction when appropriate, and estimates how many people to invite at your expected response rate.
Loading the calculator. It runs in your browser, so JavaScript needs to be on.
How to use the Sample Size Calculator
- Choose Survey proportion for a percentage, or Population mean for an average measurement.
- Select the Confidence level. Enter Margin of error in percentage points for a proportion, or measurement units for a mean.
- Enter Expected proportion, using 50% when it is unknown, or Expected standard deviation for a mean.
- Set Population size to 0 for an unknown or very large population. Otherwise enter the size of the population you will randomly sample without replacement.
- Enter Expected response rate and read both completed responses and People to invite.
The formula
For proportions, n0 = z squared times p(1-p) / e squared. For means, n0 = (z times standard deviation / e) squared. A finite population N changes this to n0 / (1 + (n0 - 1)/N). Round the result up. Invitations = completed responses / response-rate fraction, also rounded up.
Worked example
For 95% confidence, a 5 percentage point margin and p = 50%, the unrounded size is about 384.146, so you need 385 completed responses for a large population. If the whole population contains 1,000 people, the correction gives 278. At a 50% response rate, invite 556 people.
Half the margin takes four times the sample
Margin of error appears squared in the denominator. Moving from 5 to 2.5 percentage points raises the uncorrected sample requirement by a factor of four. Moving from 95% to 99% confidence also increases the sample because the normal critical value rises from about 1.960 to 2.576. These calculations assume a simple random sample.
Why 50% is the cautious starting point
The expression p(1-p) is largest at p = 0.5. If you do not know the likely survey percentage, 50% gives the largest sample requirement for the chosen margin and confidence. A reliable prior estimate of 10% gives a smaller requirement, but an optimistic guess can leave your survey underpowered for its precision target.
Completed responses are not invitations
At 25% response, a target of 400 completed questionnaires means about 1,600 invitations. That is an expected count, not a promise. Nonresponse can also introduce bias if respondents differ from everyone else. Cluster samples, weighting, repeated measurements and subgroup comparisons can require larger samples. This tool does not calculate hypothesis-test power. The mean calculation follows the known-standard-deviation precision formula.
Sample Size Calculator questions
What sample size do I need for a 95% confidence level?
You also need a margin of error and an expected proportion or standard deviation. A 5 point margin and 50% proportion require 385 responses before population correction.
When should I use population correction?
Use it for simple random sampling without replacement from a known finite population. Do not use the size of an invitation list if the target population is larger.
Is margin of error the same as confidence level?
No. Margin is the desired half-width of the interval; confidence is the long-run coverage target for the interval method.
Related calculators
- Confidence Interval CalculatorEnter a mean, standard deviation and sample size, or a success count, to calculate a t, z...
- Probability CalculatorEnter two event probabilities to find the chance of both, either, exactly one or neither....
- Standard Deviation CalculatorPaste numbers to calculate population or sample standard deviation, variance and mean....
- Z-Score CalculatorEnter a value, mean and standard deviation to find its z-score and normal percentile....
- Percentage CalculatorEnter a percent, part, whole or pair of values to find percentages and percent change....