What this Sample Size Calculator does
The Sample Size Calculator estimates how many independent observations are needed to estimate a population proportion to a chosen level of precision. It uses a normal approximation for a proportion and returns an estimated number of participants. For example, a decision analysis team might use it to plan a survey asking whether stakeholders support a proposed service change. The estimate helps make the assumptions visible before the team commits to collecting responses.
The calculator uses three inputs. The confidence level z-score represents the chosen confidence level through its corresponding standard normal value. Common two-sided choices include about 1.645 for 90% confidence, 1.96 for 95%, and 2.576 for 99%. The margin of error is the desired half-width of the estimate, entered in percentage points. The estimated population proportion is the expected share with the characteristic being measured, entered as a percent.
This is a planning model for a proportion under simple random sampling assumptions. It does not correct for a finite population, nonresponse, clustered or stratified sampling, weighting, or other design effects. It also does not calculate power for comparing groups or detecting a treatment effect. Those questions require different statistical methods and study details.
How to use the Sample Size Calculator
Choose the z-score that corresponds to your intended confidence level, then set the precision and expected proportion. If you have no defensible prior estimate for the proportion, 50% is a common conservative planning assumption because it produces the largest estimate under this formula for a fixed z-score and margin of error.
- Select a confidence level and enter its z-score. Keep the same confidence convention throughout your project.
- Set the margin of error in percentage points. A value of 5 means approximately plus or minus five percentage points under the model assumptions.
- Enter the expected proportion as a percentage from 1 to 99. Use relevant prior data or state that 50% is an assumption.
- Read the estimated participant count and round any fractional planning count upward to a whole person.
For a worked example, suppose a team chooses a 95% confidence level, uses z = 1.96, expects a proportion of 50%, and wants a margin of error of 5 percentage points. In proportion units, p = 0.50 and e = 0.05. The formula gives n = 1.96² × 0.50 × (1 − 0.50) ÷ 0.05² = 384.16. A practical target based on this estimate is at least 385 completed responses. The required number invited may be higher if some people are expected not to respond.
How the Sample Size Calculator formula works
The formula is n = z² × p × (1 − p) ÷ e². Here, n is the estimated sample size, z is the confidence level z-score, p is the expected proportion written as a decimal, and e is the margin of error written as a decimal. The calculator converts the percentage inputs to decimal proportions before applying the formula.
The expression p × (1 − p) describes the variability of a binary outcome. It reaches its maximum at p = 0.50, which is why using 50% gives a conservative estimate when the true proportion is unknown. The z-score is squared, so a higher confidence level increases the sample estimate. The margin of error is squared in the denominator. Reducing the desired error therefore increases the required sample substantially. Halving the margin of error multiplies the estimate by four, with the other inputs unchanged.
The result is an approximation based on independent observations and a normal approximation. The displayed count is not a guarantee that a particular study will achieve its intended precision. Data quality, sampling method, response patterns, and analysis decisions also matter.
Use cases for the Sample Size Calculator
In decision analysis, an estimate of a proportion can inform how much evidence a team should gather before comparing options. A community survey might estimate the share of respondents who favor one service design. A program evaluation could estimate the proportion of participants who complete a process. A planning group could measure how often a defined operational outcome occurs. In each case, the team should define the population, outcome, and sampling approach before interpreting the estimate.
The calculator is most useful when the main goal is estimating one proportion with a selected margin of error. It can support early discussion of feasibility and tradeoffs. Teams can compare scenarios by changing one assumption at a time, such as the desired precision or expected proportion, and record why each value was selected. Treat these outputs as planning estimates that guide a fuller study plan.
Other factors to consider when calculating sample size
Before using the estimate as a fieldwork target, consider whether the model matches the question and the way data will be collected. Important details include:
- Sampling design: Clustered samples and unequal selection probabilities can increase uncertainty compared with simple random sampling.
- Nonresponse: The formula estimates completed observations. The number invited may need adjustment using a response-rate assumption.
- Finite population: When sampling a substantial share of a known, bounded population, a finite population correction may be relevant.
- Group comparisons: Comparing two or more groups, or testing a difference, requires assumptions about the effect, variability, allocation, and statistical power.
- Outcome definition: Ambiguous questions or inconsistent measurement can undermine precision even when the sample count is large.
Document these assumptions alongside the estimate. If the study has a complex design or a high-stakes decision depends on the result, consult a qualified statistician or use a method tailored to the planned analysis.
Frequently asked questions
What does the margin of error mean?
It is the planned half-width of an approximate confidence interval around the estimated proportion. A 5 percentage point margin corresponds to an interval that extends about five points on either side of the estimate under the model assumptions.
Why does a 50% proportion produce a larger estimate?
The quantity p × (1 − p) is largest at 0.50. When the expected proportion is uncertain, using 50% avoids understating the estimate for this formula.
Should I round the result?
Yes. Sample counts must be whole people or observations, so round a fractional estimate upward. The example result of 384.16 becomes 385 completed responses.
Does this estimate include people who do not respond?
No. It estimates completed observations. To plan invitations, make a separate adjustment based on a justified expected response rate and explain that assumption.
Can I use this for a clinical trial or a comparison between options?
This formula estimates one population proportion and does not calculate trial power or the sample needed to compare options. Those analyses depend on the outcome, comparison, effect size, design, and other assumptions, so use a method designed for that study question.