Proportion Confidence Interval Calculator
Build a confidence interval for one population proportion with the Wald, Wilson and Agresti–Coull methods side by side.
Open → Proportion Confidence Interval CalculatorFind the sample proportion p̂ from a count and a sample size, with its standard error and margin of error. Add a known population proportion to describe the sampling distribution of p̂ and find the probability that a sample proportion falls above, below or between values.
Formula p̂ = x ⁄ n
Optional: enter a known population proportion p to describe the sampling distribution, and a value of p̂ to find how likely a sample at least that extreme is.
To calculate a sample proportion, divide the number of successes by the sample size: p̂ = x ÷ n. Its standard error is √(p̂(1 − p̂)/n), which is largest at p̂ = 0.5 and halves when n quadruples, and when np ≥ 10 and n(1 − p) ≥ 10 the sampling distribution of p̂ is approximately normal.
p̂ = x ÷ n. It is a ratio of part to whole — the same idea as the percent proportion, read as a decimal.
SE = √(p̂(1 − p̂)/n). It is largest at p̂ = 0.5 and shrinks with √n, so quadrupling n halves it.
With np ≥ 10 and n(1 − p) ≥ 10, p̂ is approximately normal, centred at p.
| Quantity | Symbol | Formula |
|---|---|---|
| Sample proportion | p̂ | x / n |
| Mean of the sampling distribution | μ_p̂ | p |
| Standard deviation of the sampling distribution | σ_p̂ | √(p(1 − p)/n) |
| Standard error (p unknown) | SE | √(p̂(1 − p̂)/n) |
| Margin of error | E | z* × SE |
These sample proportion examples cover an estimate and a probability. If 168 of 400 households own a dog, p̂ = 0.42 with a standard error of 0.0247 and a 95% margin of error of about ±4.8 points. If the true rate is 45%, a sample proportion of 0.42 or lower has about an 11% chance.
168 of 400 surveyed households own a dog. Find p̂ and its standard error.
42% ± 4.8 points
45% of households own a dog. In a sample of 400, how likely is p̂ ≤ 0.42?
About an 11% chance
Mathematically, a sample proportion is just a part-to-whole ratio — the same object as the decimal form of a ratio. What statistics adds is the question of how much that ratio would change if you drew a different sample. That variability, measured by the standard error, is what turns a single number into an estimate with honest error bars.
From here the usual next steps are a confidence interval (a range for the true proportion), a one-proportion z-test (checking a claimed value), or the sample size calculation for a future study. The guide to proportions in statistics ties these together.
p̂ = x ÷ n, read "p-hat". If 168 of 400 surveyed people own a dog, the sample proportion is 168 ÷ 400 = 0.42. It is the best single estimate of the population proportion p, which is usually unknown.SE = √(p̂(1 − p̂) ÷ n). It measures how much p̂ would vary between different random samples of the same size. If you know (or are hypothesising) the population proportion p, use p in place of p̂ — that version is the standard deviation of the sampling distribution, σ = √(p(1 − p) ÷ n).√(p(1 − p)/n), and when np ≥ 10 and n(1 − p) ≥ 10 its shape is approximately normal.