Proportion Confidence Interval Calculator

Estimate a population proportion, such as a pass rate, an approval rating or a defect rate, with a confidence interval you can trust. Enter the number of successes and the sample size (or p̂ and n). The calculator reports the Wilson score interval as the default, alongside the Clopper–Pearson exact, Agresti–Coull, Jeffreys and textbook Wald intervals, so you can compare them.

Why the Wilson interval is the default

Most textbooks teach the Wald interval, p̂ ± z√(p̂(1 − p̂)/n). It is simple but unreliable. When the proportion is near 0 or 1, or the sample is small, it can cover the true value far less often than it claims and can even go below 0 or above 1. With 1 success out of 30, for example, the Wald 95% interval is [−0.031, 0.098], which is impossible for a proportion. The Wilson interval for the same data is [0.006, 0.167]. It has close to the stated coverage in almost every situation, which is why statisticians recommend it.

Wilson: [ p̂ + z²/2n ± z√( p̂(1 − p̂)/n + z²/4n² ) ] ÷ (1 + z²/n)

Which interval should I report?

MethodBest for
Wilson scoreAlmost everything. A safe default.
Clopper–Pearson (exact)When you must guarantee at least 95% coverage, e.g. regulatory or medical work. It is wider than necessary.
Agresti–CoullTeaching: “add two successes and two failures”, then use the Wald formula.
JeffreysSmall samples. A Bayesian interval with good frequentist properties.
WaldOnly when n·p̂ and n(1 − p̂) are both at least 10, or when a course requires it.

Frequently asked questions

How do I make my interval narrower?

Increase the sample size. The width shrinks roughly with √n, so halving the margin of error needs about four times the sample. Plan it with the sample size calculator.

What if I have 0 successes?

The Wald interval collapses to [0, 0], which is clearly wrong. Wilson, Jeffreys and Clopper–Pearson still give a sensible upper limit. For 0 out of n, the 95% upper limit is roughly 3 ÷ n (the “rule of three”).