Free sample size calculator
Enter your confidence level, margin of error and expected response rate. Get the responses you need, the invitations to send and each step. No signup. All processing is in your browser.
A higher confidence level needs more responses.
The range above and below your result. For ±5%, enter 5. A smaller margin needs more responses.
The number of people in the full group. If you do not know it, leave this empty. The calculator then uses the formula for a large population.
The percentage of invited people that you expect to respond. For 20%, enter 20. If you leave this empty, you get only the responses. To measure the rate of an earlier survey, use the survey response rate calculator.
Correct this input, then calculate again:
Responses you need
To get the number of invitations, enter an expected response rate and calculate again.
Calculation steps
How many people must you invite?
A sample size formula gives the number of responses that you need. It does not give the number of people to invite. The published sample size tables in the guide by Glenn D. Israel show the number of responses that you obtain, and not necessarily the number of surveys that you send. Some invited people do not respond, so you must invite more people than the responses that you need.
This calculator gives both numbers in one result, and it shows each step. Your numbers stay in your browser.
Do you already have your responses? Use the margin of error calculator to get the margin of error of your result.
Method: Cochran's formula
For a proportion from a large population, the calculator uses the formula by Cochran:
n₀ = z² × p × (1 − p) ÷ e²
In this formula:
- z is the z-score for your confidence level.
- p is the proportion, as a decimal. This calculator always uses 0.5. This value shows the maximum variability, and it gives a conservative sample size.
- e is the margin of error, as a decimal. For example, 5% is 0.05.
This calculator uses these z-scores:
| Confidence level | z-score |
|---|---|
| 80% | 1.282 |
| 85% | 1.440 |
| 90% | 1.645 |
| 95% | 1.960 |
| 99% | 2.576 |
If you enter a population size N, the calculator applies the finite population correction:
n = n₀ ÷ (1 + (n₀ − 1) ÷ N)
The response rate is the number of people who responded, divided by the number of people that you invited. For this reason, the number of invitations is:
Invitations = responses ÷ expected response rate
The calculator rounds the responses and the invitations up to the next whole number, because a part of a response is not possible. If you round down, the result does not reach your margin of error.
This method assumes a simple random sample. A group of people who chose to take part, for example from a link on a website, is not a random sample.
Worked example
95% confidence, 5% margin of error, proportion 50%, expected response rate 20%, no population size:
- Responses for a large population: 1.96² × 0.25 ÷ 0.05² = 384.16
- Responses you need, rounded up: 385
- Invitations: 385 ÷ 0.2 = 1,925
The same input, with a population size of 1,000:
- Correction: 384.16 ÷ (1 + (384.16 − 1) ÷ 1,000) = 277.7408
- Responses you need, rounded up: 278
- Invitations: 278 ÷ 0.2 = 1,390
This is more than the 1,000 people in the population. If you invite all 1,000 people at a response rate of 20%, expect about 200 responses. To get 278 responses, you need a response rate of at least 27.8%. If you stay at 200 responses, the margin of error is about ±6.2%.
How to use the calculator
- Select the confidence level and enter the margin of error.
- If you know the size of the full group, enter the population size.
- Enter the response rate that you expect.
- Select Calculate sample size. Read the responses, the invitations and the calculation steps. If an entry is not valid, a message tells you what to correct.
Sample size FAQ
How many responses do I need for a survey?
This page does not give one fixed number, because the correct margin of error depends on the decision that you must make. For reference, at 95% confidence and a proportion of 50%, with no population size, ±10% needs 97 responses, ±5% needs 385 responses and ±3% needs 1,068 responses.
Why does the calculator use a proportion of 50%?
A proportion of 50% shows the maximum variability, so it gives a conservative sample size. The sample size can be larger than the size for the true proportion, but you do not need to know the true proportion before your survey.
What if the invitations are more than my population?
Then your expected response rate cannot give the responses that you need, even if you invite each person. The calculator shows a warning with the responses that you can expect and the response rate that you need. You can also accept a larger margin of error. For example, 200 responses from a population of 1,000 give about ±6.2% at 95% confidence.
Which response rate should I enter?
Use the response rate from an earlier survey of the same group, if you have one: the responses divided by the invitations. The calculator has no default rate, because it does not know your group. If you are not sure, calculate again with a lower rate to see how many more invitations you need.
When does the population size change the result?
It changes the result by a large amount only when the population is small. At 95% confidence and a margin of error of 5%, a population of 1,000 needs 278 responses, a population of 10,000 needs 370 responses and a population of 1,000,000 needs 385 responses.
Do more invitations make my survey more accurate?
More invitations give more responses, but they do not make the responses represent your population. A low response rate can cause nonresponse bias if the people who do not respond are different from the people who respond. The formula also does not correct a sample that is not random, for example people who chose to take part from a link on a website.
Why does the calculator say 385 and not 384?
The formula gives 384.16 at 95% confidence, a margin of error of 5% and a proportion of 50%. With 384 responses, the margin of error is a small amount more than 5%. The calculator rounds up, so the result reaches your margin of error.
Should I survey each person in a small group?
For a small population, for example 200 people or less, a census is a good option. A census removes sampling error. At 95% confidence and a margin of error of 5%, a population of 200 needs 132 responses, so at a response rate of 20% you need 660 invitations, which is more than the population.
Is my data sent to a server?
No. The calculation runs in your browser, and this page does not send or save your numbers. The site uses Plausible analytics to count page visits. Plausible does not receive what you enter.
Do I need an account?
No. The calculator is free and needs no signup.