I get back to you regarding the calculation of the 95% exact CI for the difference between 2 binomial proportions. For this matter, I would like to use the CT (Coe & Tamhane) method. Coe wrote a SAS macro for this matter (“A SAS® Macro to Calculate Exact Confidence Intervals for the Difference of Two Proportions”, which is a free article), and he explained in another article (“Small sample confidence intervals for the difference, ratio and odds ratio of two success probabilities”, which I will not be able to forward you) the steps of the algorithm for performing the 95% exact CI for the difference between 2 binomial proportions.
Does anybody have read this article? If yes, maybe you can help me… I am stuck at the Step 2.1.
I don’t understand how a min function can be used since to my mind, there is always once delta per acceptance region.
I am really embarassed because I can’t progress! I would be glad if someone could help me!!!