Dear Mr and Mrs Muthen, I wanted to ask a technical advice. When do you do a multigroup analysis, is it wrong to use a dichotomous variable as outcome? Let me explain better: in a model I wanted to consider as outcome variable the use of substances, but I don't have variables that referred to this in my data. However, I have two groups (experimental, subjects who used substances and control group, subjects who don't use substances). So, I created a new dichotomous variable where 0 = no use of substances and 1 = use of substance. Is it right to use this variable for analysis in a structural equation model?