[Feature Request]: Add unconditional exact confidence-intervals for proportion differences

[danielinteractive]

Feature description

This is described in the SAS manual here under "Exact Unconditional Confidence Limits":

• In particular we need to reproduce the RISKDIFF(METHOD=NOSCORE) option in the EXACT statement, where PROC FREQ computes the confidence limits by inverting two separate one-sided exact tests that are based on the unstandardized risk difference

• Method details - here we need to look at the "tail method" mentioned and use the unstandardized risk difference:

  

• Maybe https://github.com/cran/Exact/tree/master/R can be a good reference / starting point but we should not directly use that because there are no tests at all

• Main reference is Santner, T. J., and Snell, M. K. (1980). “Small-Sample Confidence Intervals for p 1 minus p 2 and p 1 slash p 2 in 2 times 2 Contingency Tables.” Journal of the American Statistical Association 75:386–394.

Code of Conduct

• I agree to follow this project's Code of Conduct.

Contribution Guidelines

• I agree to follow this project's Contribution Guidelines.

Security Policy

• I agree to follow this project's Security Policy.

[danielinteractive]

Example 1: group1: 12/50 events; group2: 5/45 events

Example 2: zero events in Treatment=1 (0/30), control has 3/30 events

Example 3: same but with 99% CIs