Survey sample size calculator: how many responses you need, and your margin of error

This calculator tells you how many completed responses your survey needs, how many people to invite to get them, and how precise your results are once the answers are in. The working is shown, so you can explain the number to anyone who asks.

What does a survey sample size calculator tell you?

It tells you how many answers you need before a result such as "62% agree" is close enough to what the whole workforce thinks. Two settings decide it.

• Margin of error is how far the result may sit from the true figure. At ±5%, a result of 62% means the true figure is probably between 57% and 67%.
• Confidence level is how sure you want to be that the true figure falls inside that range: 90%, 95% or 99%. Most surveys use 95%.

The target is completed responses, not invitations.

What is the survey sample size formula?

The calculator uses Cochran's formula with a finite population correction.

• Responses for a very large population: n0 = z² × 0.25 ÷ e², where e is the margin of error and z is 1.645 for 90% confidence, 1.96 for 95% and 2.576 for 99%. The 0.25 assumes an even 50/50 split of answers, the most cautious assumption, because it gives the largest sample.
• Responses for your population: n = n0 ÷ (1 + (n0 − 1) ÷ N), where N is the number of employees you invite. Round up.
• Invitations needed: n ÷ your expected response rate. Round up. It can never be more than N. If the result is larger than N, you must invite everyone, and even then you may not reach the target.
• Margin of error after the survey: with r completed responses from N people, e = z × √(0.25 ÷ r) × √((N − r) ÷ (N − 1)). This is the same formula read backwards.
• Response rate (%) = completed responses ÷ people invited × 100.

What does a worked example look like?

This example uses made-up round numbers. A company has 500 employees. It wants 95% confidence and a margin of error of ±5%, and it expects 60% of the people it invites to answer.

Other calculators may show a slightly different number for the same case, because they use a different formula. Yamane's simplified formula, used in the University of Florida's tables, gives 222.

Why do you usually invite everyone to an employee survey?

In an employee survey you rarely draw a sample. You invite everyone, which is called a census. It gives every employee a voice, and it is the only way to report results by department, site or manager. The University of Florida's guide to sample size notes that a census is attractive for populations of around 200 or fewer.

So use the calculator to know how many answers you need for the company-level result to be reliable, and, after the survey, how precise that result really is. In the example above, if all 500 employees are invited and 300 answer, the margin of error is about ±3.6%.

Team results are different. In a team of 12 where 9 people answer, the margin of error is about ±17 points. Small groups also put anonymity at risk. A common practice is to hide any group below a minimum size set before launch. Our guide on whether employee surveys are really anonymous explains how to decide and what to tell employees.

What is a good employee survey response rate?

There is no universal benchmark for employee survey response rates. The closest published reference points come from academic research, not from company surveys:

• A review of 1,014 surveys in 703 articles published in 17 journals from 2010 to 2020 found that the average response rate rose from 48% in 2005 to 68% in 2020 (Holtom and colleagues, 2022).
• An earlier review of 490 surveys in 17 journals in 2000 and 2005 found an average of 52.7% for surveys of individuals and 35.7% for surveys of organisations (Baruch and Holtom, 2008).

When we checked in November 2026, survey vendors' figures varied widely, and most did not say what they were based on. We found no published benchmark for Egypt or the Gulf that disclosed its sample. Compare your own response rate over time, by department and by site.

Does a high response rate mean the results are representative?

No. A high response rate lowers the risk of bias, but it does not guarantee that the answers represent everyone. What matters is whether the people who did not answer think differently.

The American Association for Public Opinion Research (AAPOR) notes that the response rate is no longer seen as the main measure of survey quality, and that some surveys with lower response rates have shown the least bias. Its 2016 report on survey quality summarises a 2008 meta-analysis of 59 studies by Groves and Peytcheva: the link between response rate and bias was weak, but low response rates raise the risk.

In employee surveys, the gap can be real. In a study of 194 employees, those who refused to answer an employee survey request had stronger intentions to quit, lower commitment and lower job satisfaction than those who would answer (Rogelberg and colleagues, 2000). A silent department may be your most unhappy one.

What are the most common mistakes?

• Counting invitations as responses. The target is completed answers.
• Reading a small team's result as precise. Check its margin of error first.
• Ignoring who did not answer. Compare response rates across departments, sites and languages.
• Treating the margin of error as the only error. It covers chance, not bias from who answered or from badly worded questions.

What should you do next?

Decide your margin of error and your minimum group size before launch, and write them in the survey plan. Invite everyone, track the response rate by department during the survey, and send reminders where it is low. After closing, report the margin of error next to the company result. To start with good questions, use our employee survey templates, or see how the GK Engagement Survey covers engagement, culture, EVP, experience and retention in one survey.

Frequently asked questions

How do you calculate survey sample size?

Use n0 = z² × 0.25 ÷ e², then correct it for your population with n = n0 ÷ (1 + (n0 − 1) ÷ N) and round up. For 500 employees at 95% confidence and ±5%, you need 218 completed responses.

How do you calculate a survey response rate?

Divide completed responses by the number of people invited, and multiply by 100. If you invite 364 people and 218 answer, your response rate is 59.9%.

What is the average employee engagement survey response rate?

There is no universal benchmark. Reviews of academic research surveys report averages that rose from 48% in 2005 to 68% in 2020, but these are not company engagement surveys. Compare your own rate over time.

What sample size do I need for a survey of 100 employees?

At 95% confidence and ±5%, the formula gives 80 completed responses out of 100. With so few people, invite everyone.

Sources

• Baker, R., Brick, J. M., Keeter, S., et al. (2016). Evaluating survey quality in today's complex environment. American Association for Public Opinion Research. aapor.org
• American Association for Public Opinion Research. (n.d.). Response rates: an overview. aapor.org
• Baruch, Y., and Holtom, B. C. (2008). Survey response rate levels and trends in organizational research. Human Relations, 61(8), 1139-1160. doi.org
• Groves, R. M., and Peytcheva, E. (2008). The impact of nonresponse rates on nonresponse bias: a meta-analysis. Public Opinion Quarterly, 72(2), 167-189. doi.org
• Holtom, B., Baruch, Y., Aguinis, H., and Ballinger, G. A. (2022). Survey response rates: trends and a validity assessment framework. Human Relations, 75(8), 1560-1584. doi.org
• Israel, G. D. (n.d.). Determining sample size (PEOD6). University of Florida IFAS Extension. edis.ifas.ufl.edu
• Rogelberg, S. G., Luong, A., Sederburg, M. E., and Cristol, D. S. (2000). Employee attitude surveys: examining the attitudes of noncompliant employees. Journal of Applied Psychology, 85(2), 284-293. doi.org