How to use Sample Size Calculator
- Choose Population proportion or Population mean. Set the confidence level for the precision scenario.
- For a proportion, enter the planning proportion and Margin of error (%): 5 means five percentage points, not a relative 5% error.
- For a mean, enter planning standard deviation and Margin of error in the measurement units. That margin and SD must share the same units.
- Optionally enter finite population size only for an appropriate simple random sample without replacement. Blank leaves that adjustment out.
- Calculate and review the unrounded and rounded-up counts with the assumptions. Allow separately for nonresponse, clustering or another collection design.
Example: Sample Size Calculator
Plan a proportion estimate with 95% confidence, a five-percentage-point margin and planning proportion 0.5, without finite-population adjustment.
Options
- Margin of error
- Population proportion uses Margin of error (%), where 5 means five percentage points. Population mean uses Margin of error in the measurement units, matching the planning SD.
- Planning assumptions
- Use a planning proportion or deviation appropriate to the chosen method. A finite-population adjustment is a separate assumption.
Supported inputs and limits
- Confidence level is 50.0001% to 99.999%.
- Planning proportions must be strictly between 0 and 1;
- SD and margin must be positive.
- Optional finite population is a whole number from 2 to 10¹².
- Plans whose unadjusted count exceeds one trillion are refused.
More details
This is a normal-approximation precision calculation, not hypothesis-test power, exact rare-outcome planning or estimated-SD t planning.
Nonresponse, selection bias, weighting, clustering and design effects are not added.
Finite-population adjustment assumes suitable random sampling without replacement.
A rounded count does not guarantee achieved coverage.
Where your input is processed
This tool does not upload your inputs. See the limits for anything saved on your device.
The count is for precision, not test power
The result rounds up a normal-approximation precision calculation. It does not plan power for a hypothesis test or automatically allow for nonresponse, clustering or selection bias. Keep the inputs with the count and account for the collection design separately.
Questions about Sample Size Calculator
What does a proportion of 0.5 mean here?
It supplies the largest variance for the supported proportion calculation when you have no better planning estimate.
Why is the count rounded upward?
A smaller whole count could miss the precision target under the entered model.
Does the result account for nonresponse?
No. Plan participation and sampling design separately.