Survey Sample Size Calculator
Set your confidence level and margin of error. Get the exact number of respondents your study needs — then reach them through our 30M+ verified panel.
Survey Sample Size Calculator
Leave blank for large or unknown populations.
Respondents needed
At 95% confidence, your results will be accurate to within ±5% of the true population value.
Ready to reach 385 respondents?
Buy responses from our panel →From $0.73/response · 30M+ verified respondents · 127 countries
Three inputs, one number:
Confidence level
How certain you need to be that your results reflect the real population. 95% is the research standard for most studies. Use 99% if you're making a high-stakes decision (pricing, product launch, major brand change). Use 90% only if you're doing exploratory work and speed matters more than precision.
Margin of error
How much error you're willing to accept in either direction. A ±5% margin of error means if 60% of your sample says yes, the true answer is somewhere between 55% and 65%. Tighter margins require more respondents. ±3% is common for serious research; ±5% is workable for directional studies.
Population size
Leave this blank for most studies. It only changes your sample size meaningfully when your total population is under ~20,000 people. If you're surveying all employees at a 500-person company, enter 500. If you're surveying U.S. consumers, leave it blank.
Common questions
What confidence level should I use for survey research?
95% is the standard across academic research, market research, and most professional studies. It means if you ran the same study 100 times, 95 of them would produce a result within your stated margin of error. For most brand, product, or consumer studies, 95% is the right choice. Use 99% if you're making a large financial or strategic decision based on the results — a product launch, a pricing change, a market entry. Use 90% only for quick directional checks where you already have a strong prior.
What is margin of error in survey research?
Margin of error is the range of uncertainty around your survey results. A ±5% margin of error on a 60% response means the true population answer is between 55% and 65%. It's a function of sample size — the more respondents you collect, the smaller your margin of error. Margin of error does not account for biased sampling, poor question design, or inattentive respondents. Those are quality problems, not statistical ones.
Why does the calculator assume p = 0.5?
The formula for sample size requires an estimate of the proportion you expect to find (p). Using 0.5 — the maximum variance case — produces the largest (most conservative) sample size. It's the right default when you don't know what to expect. If you already know from prior research that approximately 20% of your market does X, you can use p = 0.2 in the formula and get a smaller required sample. The calculator uses 0.5 to give you the safe upper bound.
When does population size matter?
Population size only meaningfully changes your required sample when the population is relatively small — typically under 20,000. If you're surveying a niche B2B segment with 500 potential respondents, enter that number. The finite population correction will reduce your required sample size, sometimes significantly. For mass-market B2C studies where your population is "US adults" or "UK grocery shoppers," leave it blank — the correction rounds to almost nothing.
What's the difference between sample size and response rate?
Sample size is the number of completed, qualified responses you need. Response rate is the percentage of people you contact who actually complete the study. If you need 385 responses and your expected completion rate is 20%, you need to invite approximately 1,925 people. When you buy panel responses through SegmentOS, you pay per completed response — so the 385 number is what you're buying directly.
Does a larger sample always mean better results?
More respondents reduces your margin of error — but only up to a point. Going from 200 to 400 respondents meaningfully tightens your margin. Going from 2,000 to 4,000 respondents tightens it only slightly. Beyond ~1,000–1,500 respondents, you're getting diminishing statistical returns. The bigger leverage on result quality is sample targeting (the right people), not sample size (more people). A well-screened 300-person study often produces more useful results than a poorly screened 1,000-person study.