Sample Size Fundamentals: Confidence and Margin of Error

"How many responses do we need?" has a real, calculable answer - it just depends on how sure you need to be, and how much error you can tolerate.

Start with your population

Your population is everyone you want to draw conclusions about. Sample size scales with the precision you want, not with population size, so a larger population needs only a modest additional sample once you're past a few thousand people.

Confidence level and margin of error trade off

A 95% confidence level with a 5% margin of error is the common default. Tightening either one - a higher confidence level or a smaller margin - increases the sample size you need, sometimes sharply.

Small populations get a discount

If you're surveying a known, smaller population (your whole customer base, not "the internet"), the finite population correction reduces the required sample size - our Sample Size Calculator applies it automatically once you enter your population.

Plan for non-response

If you need 400 completed responses and expect a 20% response rate, invite roughly 2,000 people, not 400 - build the shortfall into your recruitment plan from the start.

More isn't always better

Past your target sample size, additional responses add diminishing precision. Redirect that recruitment effort toward improving representativeness instead - a bigger sample that's still skewed doesn't fix a representative sample problem.