Advanced Sample Size Calculator

The calculator found below will compute the sample size given the population size, estimated population percentage, margin of error and confidence level.

If the population percentage cannot be estimated, then the Simple Sample Size Calculator should be used instead.

You may use this calculator to determine, out of a group (whose number is the population size) of people, how many should be surveyed to accurately determine the percentage who would choose a particular answer for a survey question. Since it is usually not possible to survey everyone in the entire group, only some (the sample size) are interviewed. If the sample group truly represents the whole population, then the sample percentage (which can be measured) will be no further from the actual population percentage (which is unknown) than the margin of error, with a certain confidence level.

To use this calculator, enter numbers into the boxes above the Calculate button, then click on the Calculate button to see the result underneath.

Population Size:
Population Percentage (%):
Margin Of Error (%):
Confidence Level (%):  
Sample Size:

Further information about the calculations within this calculator may be found at sample size calculations.

Observations

  • This sample size calculator requires an estimate of the population percentage, but the confidence interval calculator requires the sample percentage. Sample size calculators are used before the population is sampled, at which point the sample percentage is unknown so an estimate of population percentage must suffice.
  • Choosing a very small margin of error (to get more accurate results) can cause the sample size to grow dramatically.
  • As the population decreases below 386, the sample size grows progressively higher than half the population (assuming 95% and 5% for the confidence level and margin of error respectively).
  • As the population increases above 386, the sample size grows progressively lower than half the population (assuming 95% and 5% for the confidence level and margin of error respectively).