Summary of mi objects will output as if they were complete data.frames. Additionally, this should allow subsetting.
E.g.:
t1 <- turnout[1:1000, ]
t2 <- turnout[1001:2000, ]
t2 <- turnout[500:1000, ]
z <- zelig(vote ~ race + educate, model="logit", data=mi(t1, t2, t3))
# outputs rbind(t1, t2, t3)
summary(z)
# outputs rbind(t1, t2)
summary(z, subset = 1:2)