I’ve been running several models in which I was using kind of long labels for the columns in the DM. When I labeled columns for the interaction terms it seems that I’m using more than the allocated space because the label is truncated. My question is if I now begin running models with shorter labels will MARK know that they correspond to the same covariates that I had in previous models? Will using different labels cause problems when retrieving parameters from previous models to provide initial parameter estimates when running the new model? Will it cause problems for other procedures such as model averaging?
I’ll greatly appreciate your comments!
Andrea