This may be a bit off-topic, but as it is also about calculating Average Marginal Effects, I’ll ask it here rather than starting a new topic: Do you know how to calculate AME’s for an ordered logistic model?
I reckon the way to go is to compute the row-wise expected mean for each cutpoint (which is the probability of the ordinal score being at or below that cutpoint), and then get the AME for a particular variable using the equation you use here. In that case, there are as many AME’s as there are cutpoints (i.e. one less than the categories of your ordinal scale). However, these are still not very easy to interpret, so I was wondering whether there is a way of getting the AME for being in a particular category, rather than for being above/below a cutpoint.
Any hints, e.g. material to read up on, would be very much appreciated!