Lois Downey posted on Wednesday, August 12, 2015 - 11:24 am
I have a path model that includes 3 exogenous variables: 1 continuous and 2 dichotomous.
It also has 2 dichotomous mediating variables, and the outcome is continuous.
If I include correlations among the 3 exogenous variables, I get a message for each of the two dichotomies that the variable may be dichotomous, but is declared as continuous. However, if I declare these two variables as categorical, I get an error message, indicating that only dependent variables can be declared as categorical.
Should the correlations among dichotomous exogenous variables simply be omitted?