I have a problem with logistic regression. I want to find out if there is any effect of social capital variables (civic participation, generalized trust...) on different dependent variables (namely self-reported health, depression, cognitive functioning etc.). Similar problem had been answered here: Comparing logistic coefficients on models with different dependent variables?, but my problem is slightly different.
I had created range of different models for this dependent variables, but to achieve best fit of my models I want to add for example depression measure to my self-rated health model. I presume that being clinicaly depressed can affect self-rated health.
So my questions are: Is effect of social capital variables on this different dependent varibales comparable with slightly different independent variables lists? Is there something I can do to make my models comparable?
Thank You very much!