Survey.jl
Survey.jl copied to clipboard
Drop fpc after first stage?
Running the following in R:
> dclus2 <- svydesign(id=~dnum+snum, fpc=~fpc1+fpc2, data=apiclus2)
> dclus2_jk <- as.svrepdesign(dclus2, type="JK1", compress=FALSE)
Warning message:
In as.svrepdesign.default(dclus2, type = "JK1", compress = FALSE) :
Finite population corrections after first stage have been dropped
R drops fpc after first stage (atleast) for JK1 type designs.
We currently do the same in Julia. So even if fpc correction functionality is added, should we also keep just the first stage? Obviously investigate more how R internals work in other situations.