A binary-outcome study planner
How much does
another sample help?
Compare two sample sizes at the same assumed rate. Then use your actual counts to see how wide the interval is.
01 / Plan
Compare two sample sizes
Use the full distance between the two endpoints for your target.
| Plan | Observations | Lower | Upper | Full width |
|---|
02 / Measure
Calculate from observed counts
This section uses the confidence level selected above.
Method
Read the bounds together
The planner uses the Wilson score interval at a fixed assumed rate. The observed interval uses your actual positive and total counts. Both assume independent observations with the same underlying rate.
A confidence interval describes sampling uncertainty. Sampling bias and missing responses need their own checks.