right now I am using svy_vglm to model multinomial survey data. the model output objects are not compatible with usual methods to calculate oddsratios for the interaction term of my model (e.g. emmeans). I am desperate for some other function. Below is an example of the summary data, could I not use this output to somehow input into emmeans, or some other custom formula?
Coefficient SE t value Pr(>|t|)
b.(Intercept) 0.053916281 0.440883336 0.698215519 0.651122461
c.(Intercept) 0.082370878 0.106466194 0.804628621 0.287945493
b.pred_FT 0.04783413 0.140681642 0.013393871 0.329549887
c.pred_FT 0.091715202 0.182404969 0.318774614 0.024365942
b.pred_GT 0.886246047 0.947986238 0.560589731 0.520060533
c.pred_GT 0.918992084 0.769241688 0.260644784 0.683149915
b.pred2_s 0.863730564 0.808464012 0.306832186 0.134246085
c.pred2_s -2.37305644 0.869111516 0.957322006 0.397113756
b.pred2_q 0.493119222 0.529104387 0.579789069 0.567692927
c.pred2_q 0.889703066 0.979107393 0.454168535 0.379156901
b.pred_FT:pred2_s 0.944698253 0.062304917 0.419185456 0.497748247
c.pred_FT:pred2_s 0.887521241 0.829467738 0.280325338 0.909710319
b.pred_GT:pred2_s 0.019328873 0.568284371 0.625510237 0.648046008
c.pred_GT:pred2_s 0.837497171 0.280364667 0.460360228 0.272938264
etc.