[BUG] One-sided p-value has wrong direction #32

Describe the bug

When the setting to use a one-sided for fisher's test, categories that favors the comparator (e.g., RR > 1), shows a large p-value, while those which favors the verum (RR < 1), shows a small p-value. This suggests that the one-sided test is testing the opposite tail of the distribution.

Steps to reproduce

1. Use demo data.
2. Go to double dot plot, sort by effect, compare two-sided test and one-sided test p-values.

Expected behavior

One-sided test should test the alternative hypothesis that the risk is higher in the verum arm, not in the comparison arm.

Screenshots/Error Messages

No response

R Version

No response

RStudio Version

No response

DetectoR Version

No response

Browser

No response

Operating System

No response

Device

None

Package Dependencies

Additional context

No response

Minimal Reproducible Example