Just to double check, when I run the setx() and then the sim() command on a
dataset that is an mi()'ed version of Amelia datasets, is Zelig automatically
creating the 1/m number of simulations from each constituent dataset and making
the necessary mathematical corrections?
As in such a case:
Zelig.Imputed <- mi(out1,out2,out3,out4,out5)
ruler <- seq(0,1,0.005)
my.model <-
zelig(dprotestmar.max4~dlprotestmar.max4+polity2l+provcont+eprovcont, model =
"logit", data = Zelig.Imputed)
x.out <- setx(my.model, provcont=ruler, eprovcont=3)
s.out <- sim(my.model, x=x.out)
Thanks!
-Colin