www.cmm.bristol.ac.uk/forum

Hi,

I use Structure equation model for the Latent variable (factor) models. I came across two issues. First is, how can I put the categorical variable as predictor into the model? It is easy to put gender (male/female), but what if the variable has more than 2 categories? How can I create them for analysis?

Second, when I put tow or three variables as predictor, the results looks good, but when I put more, the results go wrong. such as the coefficient come up "1.35e-49", all the predictor are binary.

Thank you very much!

For categorical predictors use dummy variable codes. Re you results this seems to be a data problem - I suggest you look for rogue values or linear dependencies among predictors.
Harvey Goldstein