Dear all,
I have a mixed model of the form:
y[it]=a+b*z[i]+ai+(c+c[i])*x[it] ("i" refers to group and
"t" to observation in group).
zi are group-specific variables available only for a number of groups
and "ai", "ci" denote the random effects on the intercept
and slope respectively.
Is there some way to drop b*zi for the missing value cases or some other way to
deal with them?
thank you!
Irene