Results changed after assigning "ALL" to the augment types

[fanwayne]

Hello, I'm currently using this package for CMH trend analysis. However, the outputs were very confusing when I chose to set "types = 'ALL'". Below were my outputs. Could you please confirm which were correct? Many thanks.

vcdExtra::CMHtest(Freq ~ Var2 + Var1, data = table_4cmh) Cochran-Mantel-Haenszel Statistics for Var2 by Var1

             AltHypothesis   Chisq Df     Prob

cor Nonzero correlation 3.3314 1 0.067967 rmeans Row mean scores differ 5.7281 4 0.220397 cmeans Col mean scores differ 4.8995 4 0.297764 general General association 14.5570 16 0.557290

vcdExtra::CMHtest(Freq ~ Var2 + Var1, data = table_4cmh, types = "ALL") Cochran-Mantel-Haenszel Statistics for Var2 by Var1

             AltHypothesis   Chisq Df     Prob

general General association 3.3314 1 0.067967 rmeans Row mean scores differ 5.7281 4 0.220397 cmeans Col mean scores differ 4.8995 4 0.297764 cor Nonzero correlation 14.5570 16 0.557290