This one proportion z test calculator compares a sample proportion with a hypothesized population proportion. Enter the number of successes, the sample size, the null value p₀, the direction of the alternative hypothesis, and α. The result reports p̂, the null standard error, z statistic, p-value, decision, large-count check, and every substitution used in the calculation.

Use this test for one binary outcome measured in one independent sample—for example, yes/no, converted/not converted, or defective/not defective. It is a large-sample normal approximation, not an exact binomial test.

02Input

How to Use the One Proportion Z Test Calculator

Define a success before looking at the result, then enter the number of successes x and the total sample size n.
Enter the claimed population proportion as p₀. Use a decimal such as 0.50, not 50 for 50 percent.
Choose a two-sided, greater-than, or less-than alternative. The direction must come from the research question, not from which direction the sample happened to move.
Enter the significance level α. A value of 0.05 is common, but the appropriate threshold should be chosen before testing.
Run the test and read the p-value, practical context, and condition check together. A small p-value does not repair a weak sampling design.

The calculator accepts whole-number counts with 0 ≤ x ≤ n, requires 0 < p₀ < 1, and requires 0 < α < 1. It does not accept percentages with a percent sign.

03Calculation

One Proportion Z Test Calculator Formula

First calculate the observed sample proportion:

p̂ = x / n

Under the null hypothesis H₀: p = p₀, the standard error is based on the null value rather than the observed proportion:

SE₀ = √[p₀(1 − p₀) / n]

The one proportion z test statistic is:

z = (p̂ − p₀) / SE₀

The z statistic measures how far the observed proportion is from the null value in null-standard-error units. The selected tail converts that z statistic to a p-value. For a two-sided test, the calculator doubles the standard-normal tail beyond |z|. For a greater-than test it uses the area to the right of z; for a less-than test it uses the area to the left.

Decision ruleP ≤ α

When the p-value is at or below alpha, the data provide enough evidence against the null claim for the chosen alternative. When it is above alpha, the data do not provide enough evidence; that result does not prove the null is true.

04Direction

Choose the Correct One Proportion Test Hypotheses

Every one-sample proportion test here uses H₀: p = p₀. The research question determines the alternative and therefore which part of the z distribution counts as more extreme.

ATwo-sidedp ≠ p₀

Is the Population Proportion Different?

Choose two-sided when either an increase or a decrease would answer the question. Evidence in both tails contributes to the p-value.

BRight-tailedp > p₀

Is the Population Proportion Greater?

Choose greater when only values above p₀ support the stated alternative. A negative z statistic will not support this direction.

CLeft-tailedp < p₀

Is the Population Proportion Less?

Choose less when only values below p₀ support the alternative. Do not switch to this tail after seeing a low sample proportion.

05Worked example

One Proportion Z Test Example with P-Value

Suppose a team claims that half of all eligible users complete an onboarding step. In a preplanned random sample, 62 of 100 users complete it. Test whether the population completion proportion differs from 0.50 at α = 0.05.

The hypotheses are H₀: p = 0.50 and Hₐ: p ≠ 0.50.

Sample proportion: p̂ = 62 / 100 = 0.62.
Null standard error: SE₀ = √[0.50(1 − 0.50) / 100] = 0.05.
Test statistic: z = (0.62 − 0.50) / 0.05 = 2.40.
Two-sided p-value: P ≈ 0.016395.
Because 0.016395 ≤ 0.05, the sample provides enough evidence that the population completion proportion differs from 0.50.

The condition calculation uses the null value: np₀ = 50 expected successes and n(1 − p₀) = 50 expected failures. Both exceed the calculator’s threshold of 10, so the large-count check passes. The observed difference is 12 percentage points; whether that difference matters in practice is a separate question from statistical significance.

The observed proportion is 2.4 null standard errors above the claim; two-sided P ≈ 0.0164.

06Validity

Conditions and Assumptions for a 1 Proportion Z Test

The arithmetic is valid only when the study supports the model. Review all of these conditions before using the conclusion:

Binary outcome: every observation must be classified consistently as a success or failure.
Representative data: a random sample or defensible randomization process is needed for inference to the stated population. Convenience data can be precisely calculated and still be biased.
Independent observations: one observation should not determine another. If sampling without replacement from a finite population, the sample is commonly kept to no more than about 10 percent of that population unless a finite-population method is used.
Large counts under the null: this calculator flags whether np₀ ≥ 10 and n(1 − p₀) ≥ 10. When either count is smaller, an exact binomial procedure may be more appropriate.
Preselected direction and alpha: choosing the tail or significance level after inspecting the result changes the error rate and makes the reported p-value misleading.

Only the large-count condition can be checked from the entered totals. The calculator cannot determine whether sampling was random, observations were independent, missing data were informative, or success was measured reliably.

07Interpretation

How to Interpret the Z Score, P-Value, and Decision

Does a p-value give the probability that the null hypothesis is true?

No. The p-value is calculated assuming the null model is true. It describes how extreme the observed z statistic, or a more extreme one in the selected direction, would be under that model. It is not a posterior probability for H₀.

What does “not enough evidence” mean?

It means the data did not cross the selected evidence threshold. It does not show that p equals p₀, and it may reflect a small effect, limited sample size, high uncertainty, or a true null value. Equivalence requires a test and margin designed for equivalence.

Is the sample proportion the same as the population proportion?

No. p̂ is the observed statistic; p is the unknown population parameter. The z test evaluates one claim about p. To estimate a range of plausible values, use a proportion confidence interval calculator.

Does statistical significance mean the effect is important?

No. Report the observed difference p̂ − p₀ and its real units alongside the p-value. A tiny difference can be statistically detectable in a very large sample, while an important difference can remain uncertain in a small sample.

08Method

Method, Limitations, and Statistical Sources

This calculator uses the uncorrected large-sample normal approximation. It does not calculate an exact binomial p-value, apply a continuity correction, adjust for multiple testing, model clustered or paired observations, handle survey weights, or infer causation. Results may therefore differ from software using an exact or continuity-corrected method. Keep the method consistent when comparing outputs.

The formulas and interpretation were checked against the NIST Engineering Statistics Handbook section on testing a proportion and Penn State STAT 200’s one-proportion hypothesis-testing lesson. Penn State’s np₀ ≥ 10 and n(1 − p₀) ≥ 10 guideline is the threshold reported by this calculator. These educational references do not replace a statistical analysis plan or subject-matter review for high-stakes decisions.