Sample Size Calculator for Surveys and Market Research
Work out how many survey completes you need for any margin of error and confidence level, with finite population, design effect, subgroups and invites. Free.
inquiry@globalvoxpopuli.com · globalvoxpopuli.com
How to use the sample size calculator
- Choose what you are measuring. Pick "A percentage" for questions like awareness, purchase intent or agreement. Pick "An average" for numeric answers like spend, ratings or minutes.
- Set your confidence level. 95% is the market research standard. Use 90% for directional work and 99% when a wrong call is expensive.
- Set your margin of error. ±5 points is common for general studies, ±3 for trackers and published polls, and ±7 to ±10 for hard-to-reach B2B or physician audiences.
- Enter the expected result if you have one. Leave 50% if you don't; it gives the largest, safest sample.
- Add the population size only for small, countable groups such as oncologists in Belgium or IT directors at listed companies.
- Use the advanced fields when you plan to weight the data or report subgroups, and the fieldwork fields to see how many people you need to invite.
Every input is saved in the page address. Use Share link to send the exact scenario to a colleague, or CSV to drop the results into a proposal.
What the result means
The main number is the count of completed interviews you need for the precision you chose. If the calculator says 385 completes at ±5 points and 95% confidence, and 40% of your sample say they would buy, the true figure for the population is very likely between 35% and 45%. "Very likely" means that if you ran the same survey 100 times, about 95 of those intervals would contain the true value.
The margin of error only covers random sampling error. It does not cover a biased sample, leading questions, poor translation or fraudulent respondents. Those are fixed by sample sourcing, questionnaire design and quality control, not by a bigger sample.
The sample size formula
For a percentage, the calculator uses the standard formula for estimating a proportion:
n = z² × p × (1 − p) ÷ E²
where z is the z-score for your confidence level (1.96 for 95%), p is the expected proportion and E is the margin of error as a decimal. For an average it uses n = z² × σ² ÷ E², where σ is the standard deviation.
Two corrections follow. The design effect inflates the sample to make up for weighting or clustering, which make each interview worth less than one in a simple random sample. The finite population correction, n ÷ (1 + (n − 1) ÷ N), reduces the sample when you are surveying a large share of a small population.
Worked example
A brand team wants purchase intent in three countries, reported separately, accurate to ±4 points at 95% confidence. Last year's intent was about 30%. They will weight to census age and gender, which they expect to cost a design effect of 1.3.
- Base sample: 1.96² × 0.30 × 0.70 ÷ 0.04² = 504.2
- With the design effect: 504.2 × 1.3 = 655.5, rounded up to 656 per country
- Total: 656 × 3 = 1,968 completes
At 35% incidence, an 18% response rate and 6% removed in quality checks, they need about 33,200 invitations. Press Example above to load this scenario.
Sample sizes for international and multi-country studies
In a multi-country study, each country is a subgroup. If you need to compare countries, each one needs its own robust base, so total sample grows with every market you add. If you only need a global total, you can allocate unevenly and weight back, but the design effect will rise. Our sample allocation calculator compares proportional, equal and compromise splits and shows the precision of each.
Hard-to-reach audiences change the maths in practice. A country may have only a few hundred cardiologists who take part in research, so the finite population correction matters, and feasibility, not the formula, often sets the ceiling. The feasibility checker checks your target against panel sizes by country.
Quick reference: common sample sizes
| Margin of error | 90% confidence | 95% confidence | 99% confidence |
|---|---|---|---|
| ±10 pts | 68 | 97 | 166 |
| ±7 pts | 139 | 196 | 339 |
| ±5 pts | 271 | 385 | 664 |
| ±4 pts | 423 | 601 | 1,037 |
| ±3 pts | 752 | 1,068 | 1,844 |
| ±2 pts | 1,691 | 2,401 | 4,147 |
Figures assume p = 50%, a large population and no design effect.
Frequently asked questions
What is a good sample size for a survey?
There is no single number. For most consumer surveys, 400 completes per group you want to report gives about ±5 points at 95% confidence. Trackers and published polls often use 1,000 or more (about ±3 points). B2B and physician studies often accept 100 to 200 per segment because the population is small and costly to reach.
Does population size matter?
Only when it is small. The sample needed for ±5 points is 385 for a population of 100,000 and 384 for a population of 100 million. It drops noticeably only when your sample would be more than about 5% of the population.
Why use 50% if I don't know the expected result?
The term p × (1 − p) is largest at p = 0.5, so 50% gives the biggest sample. Any other true value will then be measured at least as precisely as you planned.
What is a design effect?
It is the ratio of the variance of your actual design to that of a simple random sample of the same size. Weighting, clustering and quota designs usually push it above 1. A design effect of 1.5 means you need 50% more interviews for the same precision. The weighting efficiency calculator estimates it from your weights.
Can I use this for online panel samples?
Yes, as a guide. Margin of error formulas assume probability sampling, and online panels are non-probability samples. Industry bodies such as AAPOR suggest calling the figure a "credibility interval" or describing precision carefully. The sample size logic still holds for planning.
How many completes do I need per subgroup?
Treat each subgroup you will read on its own as a separate sample. Many researchers set a floor of 100 per subgroup (±10 points) and avoid reporting cells under 50.
Need the respondents, not just the numbers?
Global Vox Populi runs quantitative and qualitative fieldwork in 170+ countries through its own proprietary panels of consumers, B2B and IT decision-makers, physicians, nurses, patients, caregivers and payers. ISO 9001, ISO 20252 and ISO/IEC 27001 certified and HIPAA compliant, following ESOMAR and Insights Association guidelines.
Check feasibilityLink to this tool
Free to use in proposals, courses and articles. If it helped, a link back is appreciated:
<a href="https://globalvoxpopuli.com/tools/sample-size-calculator/">Sample Size Calculator</a> by Global Vox Populi