# Sample size calculator > How many responses does your survey need? The number of survey responses you need for a margin of error and confidence level, for any population size. Find how many survey responses you need for a chosen margin of error and confidence level, with the finite population correction for small groups. - Page: https://how.box/sample-size/ - Runs entirely in the browser. - License: HowToolbox's own text and data (titles, explanations, FAQs, tool descriptions and schemas, this file) are CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/); AI training and answers are welcome. Attribute "HowToolbox (how.box)" with a link to the page used. Third-party data keeps its own terms; the name and logo are not licensed. Publisher: HowToolbox, contact@how.box. Terms: https://how.box/about/terms/ ## Inputs Answer link: https://how.box/sample-size/?e={e}&c={c}&pop={pop}&n={n}&p={p} - `e` (number 0.1 to 50): Margin of error, ± percentage points. Example: `5`. - `c` (number 50 to 99.99): Confidence level in percent. Example: `95`. - `pop` (large or known): large (default) or known. Example: `known`. - `n` (whole number 1 to 1e9): Population size (pop=known). Example: `1000`. - `p` (number 1 to 99): Expected share in percent. Example: `50`. Example answer: https://how.box/sample-size/?e=5&c=95&pop=known&n=1000&p=50 A complete link carries the primary answer in the page HTML; a missing or invalid value shows the default page. ## Agent tools - `sample_size_calculate`: Survey sample size to estimate a proportion within ± margin_percent points at confidence_percent confidence: n₀ = z² p(1 − p) / e² (p = proportion_percent ÷ 100, 50 by default and the most cautious), rounded up. - MCP, one POST to https://how.box/mcp: `{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"sample_size_calculate","arguments":{"margin_percent":5,"confidence_percent":95,"population":"known","population_size":1000,"proportion_percent":50}}}` - GET: https://how.box/mcp/run/sample_size_calculate?margin_percent=5&confidence_percent=95&population=known&population_size=1000&proportion_percent=50 ## How it works A survey of a sample estimates what the whole group thinks. The margin of error is how far the sample's answer may be from the true share, and the confidence level is how often a range that wide would catch the truth if you repeated the survey. Narrower margins and higher confidence both need more responses. The basic formula is n₀ = z² × p(1 − p) ÷ e², where z is the critical value of the normal distribution for the confidence level (1.96 for 95 %), p the expected share and e the margin as a fraction. For ±5 % at 95 % with p = 0.5 that is 1.96² × 0.25 ÷ 0.05² ≈ 384.2, so 385 responses. For a known population N the finite population correction n = n₀ ÷ (1 + (n₀ − 1) ÷ N) lowers it: 278 for a population of 1,000. The formulas follow the NIST/SEMATECH e-Handbook and Cochran's Sampling Techniques. The result counts completed responses from a random sample. It cannot fix a biased sample: people who choose to answer may differ from those who don't, so plan to invite more people than the number shown. An expected share of 50 % gives the largest, safest size. ## FAQ **Why does a population of a million need almost the same sample as a population of 10,000?** Precision depends mostly on how many people answer, not on what fraction of the population they are. The correction matters only when the sample is a sizeable part of the whole group. **What confidence level should I use?** 95 % is the common choice. 90 % needs fewer responses, and 99 % needs about 70 % more than 95 %. **What should I put as the expected share?** Leave it at 50 % unless an earlier survey tells you the answer will be far from even. 50 % gives the largest sample, so it is never too small. **Does this tell me how many invitations to send?** No, it gives the responses you need. Divide by the response rate you expect: needing 385 responses at a 20 % response rate means inviting about 1,925 people. **Can I use this to estimate an average?** No. This calculator estimates a proportion, such as the share answering yes. Planning a sample for an average also needs an estimate of the values’ spread. More: [HowToolbox index](https://how.box/llms.txt) · [Agent Skill](https://how.box/skill.md) · [agent terms](https://how.box/about/agents/)