Statistical Significance Calculator

Check whether a control vs variant difference is statistically significant.

Control group
Treatment / variant group

Control 20% vs variant 25%

Statistically significant at α = 0.05

Observed difference
-5%
z statistic
-2.6774
p-value
0.00742

Control rate: 20% · Variant rate: 25%

Statistical significance does not prove causation or practical importance.

How statistical significance works

This calculator compares two proportions — for example a control group and a treatment or variant group in an A/B test — using a two-proportion z-test.

It reports the observed difference, the pooled standard error, the z statistic, and the two-tailed p-value. If the p-value is below your chosen significance level, the difference is statistically significant.

Statistical significance only tells you that the difference is unlikely to be due to random chance. It does not prove that one option caused the outcome, and a statistically significant result can still be practically unimportant.

The formula (pooled two-proportion z-test)

z = (p̂₁ − p̂₂) ÷ √(p̂(1−p̂)(1/n₁ + 1/n₂)) · p̂ = (c₁+c₂) ÷ (n₁+n₂)

p̂₁ and p̂₂ are the two observed conversion rates, p̂ is the pooled proportion under the null hypothesis, and the z statistic is compared to the standard normal distribution to get the two-tailed p-value.

Worked example

Control 200/1000 vs variant 250/1000

Rates are 20% and 25%. The pooled proportion is 22.5%, giving z ≈ 2.68 and a two-tailed p-value of about 0.007 — statistically significant at the 5% level.

Frequently asked questions

What assumptions does the test make?

Independent random samples and an adequate normal approximation (expected counts of at least 5 in each cell). Small samples can make the approximation unreliable.

What is the difference between statistical and practical significance?

Statistical significance means the result is unlikely under chance. Practical significance asks whether the size of the effect matters in the real world — a tiny but statistically significant difference may not be worth acting on.

Does a significant result prove causation?

No. An observational difference or even a controlled experiment only supports causation under strong design assumptions. This calculator only reports the statistical comparison.

Built by PanelRoster

PanelRoster is an enterprise operations management platform for organizations coordinating distributed teams, assignments, workflows, quality assurance and operational execution.

Explore PanelRoster