Hello guys, I am using MICE approach. I have some continuous variables (e.g., income) and categorical variables (e.g., race as dichotomous) and dependent variables like cogtest 1, cogtest2. For moderation, I need to use several interaction terms. Most of the independent and all dependent variables have missing values except 2 independent variables gender and age . I am familiar with treating interaction terms as 'just another variable' approach. i also know about them keeping them as passive. Now the question is: should I include them as new variables (e.g., after generating one (incrace= income X race) ), If I do should it be REGRESS or LOGIT? Should it be something like MI Imputed Chained (Regress) income incrace cogtest1 cogtest2 (logit) race= Gender Age, add (10)...? It doesnt make much sense as race has not been imputed yet. Should I have 2 commands then, the first one without interaction terms and the second equation command include all the interaction terms? or should I just include them as passive? It might sound silly but I am pretty confused as I am just learning.