[R] Multinomial logistic regression in an ANOVA-like framework

[Stats Wolf]

Dear all,

I have a problem for multinomial logistic regression: the response
variable is multinomial (score 1-5) and the two predictors are
categorical; all that comes from panelists (it's a kind of preference
study), which I treat as a block and include in the model (is it
correct?). I would like to see the results in the ANOVA-like
framework. Fitting the multinom() function from the nnet package works
pretty well and using the Anova() from the package car I got the
hypothesis testing for the predictors as well as their interaction.
However, I would like to go further and (because I am do look at means
when interpreting the results) I would like to make some kind of
multiple comparisons, if it is at all possible.

Any suggestions if I am doing anything wrong and if not, what could I
do to get these multiple comparisons?

Thank you,

Stats Wolf