How to Calculate the Sample Size Your Survey Actually Needs

You’re designing a survey or planning an experiment and need to know how many responses actually make the results meaningful — too few and your findings could easily be noise; collecting far more than necessary wastes time and resources on responses that don’t meaningfully improve your result’s precision. Sample size isn’t a number you should guess at — it follows directly from a few statistical inputs that determine exactly how many responses you need to hit a specific confidence level and margin of error.

Getting this number wrong in either direction has real costs: an underpowered survey risks drawing conclusions that don’t actually hold up, while an oversized one burns budget and respondent goodwill collecting data beyond what your precision requirements actually call for.

What determines required sample size

Four inputs drive the calculation: population size (how many people could possibly respond, which matters less as it gets very large), confidence level (how certain you want to be that your results reflect the true population — commonly 95%), margin of error (how much wiggle room you’re willing to accept in your results — commonly ±5%, tighter margins requiring more responses), and expected response split (how evenly you expect answers to divide, with a 50/50 split requiring the largest sample size for a given confidence and margin, since that’s the most statistically “uncertain” scenario).

Tightening the margin of error or raising the confidence level both increase the required sample size, and the relationship isn’t linear — cutting your margin of error in half, for instance, roughly quadruples the required sample, which is exactly why an intuitive guess at “enough responses” tends to be badly wrong in one direction or the other.

Why people get stuck here

• Guessing at a “reasonable” sample size instead of calculating it. A round number like 100 or 500 responses feels sufficient but has no actual connection to the confidence and precision the survey needs.
• Not accounting for population size correctly. For very large populations, required sample size levels off and doesn’t keep scaling proportionally, which is counterintuitive if you’re extrapolating from a smaller population’s requirements.
• Underestimating how much tightening the margin of error costs. The nonlinear relationship between margin of error and required sample size means a seemingly modest precision improvement can require a dramatically larger sample.
• Ignoring expected response split. Assuming a lopsided expected split (like 90/10) when the actual answer is closer to 50/50 underestimates the sample size actually needed for reliable results.

What a good sample size calculator looks like

Takes all four relevant inputs

Population size, confidence level, margin of error, and expected response split all genuinely affect the calculation, so a calculator needs all four to produce an accurate result rather than a rough approximation.

Explains the trade-offs clearly

Since tightening precision or raising confidence increases required sample size non-linearly, showing how the result changes as you adjust these inputs helps you make an informed trade-off between precision and practicality.

Produces a clear minimum sample size

The output should be a specific, actionable number — the minimum responses needed to meet your stated confidence and margin of error — not just a vague range.

Common mistakes to avoid

• Choosing a round sample size number without any actual statistical basis for why it’s sufficient for your specific confidence and precision needs.
• Assuming required sample size keeps scaling proportionally with population size even for very large populations, where the relationship actually levels off.
• Underestimating the sample size cost of a tighter margin of error, given the nonlinear relationship between the two.
• Assuming a lopsided expected response split when reality is closer to 50/50, understating the required sample size.
• Treating a calculated sample size as a hard guarantee of accuracy — sampling assumptions (like random selection) also need to hold for the statistical guarantees to actually apply.

How to do it with Sample Size Calculator

Online Tool Store’s Sample Size Calculator runs entirely in your browser.

1. Open the Sample Size Calculator tool.
2. Enter your population size, target confidence level, and acceptable margin of error.
3. Set your expected response split.
4. Review the minimum sample size needed to meet your stated requirements.

Because it’s instant, you can adjust confidence level and margin of error to see the trade-off before finalizing your survey plan.

Frequently asked questions

What confidence level and margin of error should I use?

95% confidence with a ±5% margin of error is a common default for general surveys, striking a reasonable balance between statistical rigor and practical sample size — but higher-stakes research might reasonably require tighter margins or higher confidence, at the cost of a larger required sample.

Does a bigger population always require a bigger sample size?

Not proportionally — as population size grows very large, the required sample size for a given confidence and margin of error levels off rather than continuing to scale up, which is why a national survey doesn’t need dramatically more responses than a much smaller regional one at the same confidence and precision.

What if I can’t reach the calculated minimum sample size?

You can still run the survey with fewer responses, but your results will have a wider actual margin of error (or lower confidence) than your original target — it’s worth being transparent about that limitation when reporting results based on a smaller-than-ideal sample.

Final thought

Sample size isn’t a number to guess at — it’s determined by the confidence and precision you actually need, and the relationship between them is steep enough that intuition alone reliably gets it wrong.

Try the free Sample Size Calculator tool