When using the plus four confidence interval for the difference p₁ - p₂, what do we add?

Question

Quizplus · Accepted Answer

To use the plus four confidence interval, we add one success and one failure to each of the two samples. This adjusts the interval to be slightly wider to account for the added uncertainty due to the small sample sizes. Adding two successes or two failures would not be appropriate as it would change the underlying proportions and skew the data. Adding four outcomes to the smallest sample would also not be appropriate as it would disproportionately affect the interval width and not reflect the true uncertainty in the data.

When Using the Plus Four Confidence Interval for the Difference