Hi, I have a question about using adjust after running Logit model with categorical variables as controls. For example, k, m and n are all categorical variables, with 4, 2, 3 categories for each separately. I have a model as:

.logit y i.k m#n X

I need to predict the probability at means of all the categorical variables, by(X), so I used "adjust" post estimation command. However, I got the error "operator invalid", which I know means I shouldn't have used factor variables and interactions in the logit model.

But how shall I get around this problem? Please advise.

Many thanks!