Is it wrong to use margins command for un-centered variables when my model uses centered versions?

[Deepali godha]

Hi All,

I have conducted a two level logistic model using

svy, subpop():melogit

My level 1 variables are group-mean centered and level 2 variables are grand mean centered. But instead of using the grand mean centered variables, I want to use the cluster means to make my graphs easier to read. I had put a similar problem in an earlier post here but later decided to stick to the model form itself. This time around, I need to find a way out but I am not sure. As you will notice, using un-centered variables in that post had led to change in interpretation. I will appreciate any guidance.

Thank you
Deepali