Calculator
Example Data Table
| Scenario | Group 1 successes | Group 1 size | Group 2 successes | Group 2 size | Test direction |
|---|---|---|---|---|---|
| Email conversion test | 56 | 400 | 45 | 350 | Two sided |
| Defect rate comparison | 18 | 250 | 31 | 280 | Less than |
| Survey support rate | 144 | 600 | 118 | 580 | Greater than |
Formula Used
Sample proportions:
p1 = x1 / n1
p2 = x2 / n2
Pooled proportion for a zero difference null test:
p = (x1 + x2) / (n1 + n2)
Z statistic with pooled standard error:
z = ((p1 - p2) - d0) / sqrt(p(1 - p)(1 / n1 + 1 / n2))
Unpooled standard error:
SE = sqrt(p1(1 - p1) / n1 + p2(1 - p2) / n2)
Confidence interval:
(p1 - p2) ± z critical × SE
How to Use This Calculator
Enter the success count and sample size for both groups. Select the alternative hypothesis. Add alpha and confidence level values. Use the pooled option for a standard zero difference test. Use the continuity correction when you want a more conservative result for smaller samples. Press Calculate to view the result above the form.
Article
Why Two Proportion Testing Matters
A two proportion z test compares rates from two independent groups. It helps when the result is a yes or no outcome. Common examples include conversions, defect rates, disease rates, pass rates, and survey choices. The calculator turns raw counts into proportions, then measures whether the observed difference is large enough to be unlikely by chance.
Good testing starts with clean inputs. Each group needs a sample size and a success count. The success count must not exceed the sample size. Samples should be independent. Each observation should belong to only one group. The normal approximation works best when each group has enough expected successes and failures.
What the Result Means
The z score shows how many standard errors separate the observed difference from the null difference. A large positive z score favors a higher first proportion. A large negative z score favors a higher second proportion. The p value converts that distance into probability under the null hypothesis.
The decision depends on alpha. If the p value is less than alpha, the test rejects the null hypothesis. This does not prove a practical effect. It only shows statistical evidence. Always compare the confidence interval and real-world cost before acting.
Advanced Options
The pooled standard error is common when the null difference is zero. It assumes both groups share one underlying proportion during the test. The unpooled standard error uses each group separately. It is useful for confidence intervals and nonzero null differences. The continuity correction can make the test more conservative for smaller samples.
The calculator also reports the confidence interval for the difference. This interval gives a range of likely values for p one minus p two. If the interval includes zero, the difference may not be statistically clear at the chosen confidence level.
Practical Notes
Use two sided testing when any difference matters. Use right tailed testing when the first group must be higher. Use left tailed testing when the first group must be lower. Report the sample counts, proportions, z score, p value, confidence interval, and test direction. This creates a transparent result that readers can check. Keep records, because later audits often need the original assumptions and labels for review.
FAQs
What is a two proportion z test?
It is a hypothesis test that compares two independent sample proportions. It checks whether their observed difference is statistically meaningful.
When should I use this calculator?
Use it for binary outcomes. Examples include yes or no answers, pass or fail results, defect rates, and conversion rates.
What does the p value mean?
The p value shows how likely the observed difference is under the null hypothesis. Smaller values give stronger statistical evidence.
What is the pooled proportion?
The pooled proportion combines successes and sample sizes from both groups. It is often used when the null difference is zero.
Should I use a two sided test?
Use a two sided test when either group could be higher. It is the safest choice when you do not have a fixed direction.
What is alpha?
Alpha is the significance cutoff. A common value is 0.05. If the p value is lower, the result is statistically significant.
What does the confidence interval show?
It gives a likely range for the true difference between the two population proportions. A wider interval means more uncertainty.
Can sample sizes be different?
Yes. The calculator supports unequal sample sizes. Each group only needs its own valid success count and total sample size.