Survey Sample Size Calculator

Survey sample size is the number of completed respondent answers needed to estimate a result with a chosen confidence level and margin of error. In research, it helps determine how much data you need from the target population or audience segment before running analysis. It is not the number of invitations sent, because not everyone invited to the questionnaire will respond.

The sample size formula uses p(1 - p), which is largest when the expected proportion is 50%. That means the population is assumed to be most evenly split on the survey question, so the calculator recommends the most conservative estimate. This is useful for a new market poll or study when you do not have prior data.

Only up to a point. Once the population is large, the required sample size changes very little because the margin of error is driven more by the sample than by the total audience. Population size matters most when your sampling frame is a small, defined group such as a class, employee list, customer segment, or member database.

A 5% margin of error is common for many general surveys because it gives a practical balance between accuracy and sample size. Use a smaller error margin when precision matters, such as comparing close poll results or evaluating a high-stakes market decision. Expect the required number of completed responses to increase quickly as the margin gets smaller.

No. Sample size controls statistical precision, but accuracy also depends on sampling methodology, response bias, question wording, and whether respondents represent the audience you want to measure. A large dataset can still produce a weak statistic if representativeness is poor or the survey design pushes people toward a certain answer.

Completed responses are the usable survey answers you need for the final analysis. Invitations are the number of people you may need to contact to get that many responses, based on your expected response rate. For example, if the calculator says you need 400 completed responses and your expected rate is 25%, you may need to invite about 1,600 people.

Use a 95% confidence level for most general surveys because it is a widely accepted balance between reliability and sample size. A 90% confidence level needs fewer responses, while 99% confidence is more conservative and requires more completed surveys. The best confidence setting depends on how much risk your study can tolerate.

Enter a population size when you know the total audience, such as 800 employees, 2,500 customers, or 12,000 members. The calculator can then apply a finite population correction, which may reduce the required sample. If the audience is very large or unknown, leaving the field blank is usually appropriate.

This calculator is best for proportions and percentages, such as the share of respondents who choose an option in a questionnaire. Rating scale averages can use different sample size methods, especially when you care about the mean score rather than a percentage. If your analysis is based on a yes/no result or a selected option, this proportion-based formula is usually a good fit.

If the response rate is lower than expected, you may need to send more invitations, extend the survey window, or use reminders. The completed response target does not change, but the outreach effort needed to reach it increases. Low response can also affect representativeness if one audience segment is much less likely to answer than another.