Dear all,
I have the following question:
I estimate a model of the form
xi: logit y x1 i.x2
Then I attempt to do the below:
prvalue, x(x2=0) rest(mean)
prvalue, x(x2=1) rest(mean)
But I get an error message:
x1 not rhs variable
invalid syntax
r(198);
Any ideas? Is it that I can't use the xi route for
specifying categorical rhs's and then use prvalue?
x2 isn't actually a variable in the equation. Dummies created from
x2 are. So, I think you need to create the dummies
yourself. Although, if x2 is already a dummy var, you don't need xi
in the first place.