Added useful contributions by Howard on a related approach. Method now also offers a permutation test equivalent method, plus an interpretable statement of effect as proposed by Howard.
Patched a bug when switching between dropping and splitting ties. Previously, when assuming null standard error and dropping ties confidence intervals and p-values would erroneously return NA values.
Added option for controlling standard error. The use of standard error under Null for calculations (Following O’Brien’s approach is now achieved by means of the assume_no_effect option).
Patched a bug in the calculation of cell proportions for the purposes of calculating standard error under null. Previous version assumed equal group sizes.
NOTE: I forgot to upload this to Github for the dev version. See cran/genodds for this version.
Package released