Question about modelling ordinal predictor and response variables

Hi all

I am new to modelling with ordinal variables and am unsure if I understand their specification correctly, both when used as predictor and response variables.

Say I have a normally distributed continuous variable, con_var, and a correlated ordered categorical variable, ordinal_var, and I am interested in predicting either of them, given the other. Would I be correct in understanding that the two example models below would be a reasonable way to model their relationship, and that they both match the basic assumptions given above - or am I missing something?

model1 = ordinal_var ~ con_var, family = cumulative()

model2 = con_var ~ mo(ordinal_var), family = gaussian()

Thanks!

1 Like

This looks like a fine place to start.