I would like to apply a statistical significance test to a contingency table where the rows are an ordinal variable and the columns are a nominal variable.
Statsmodels implements tests for nominal/nominal (chi-squared) and ordinal/ordinal (linear-by-linear) contingency tables.
From my research, the appropriate test for ordinal/nominal appears to be an Extended Cochran-Armitage test [EDIT: Originally had Cochran-Armitage].
How can I apply an extended Cochran-Armitage in Python?