Survey.jl
Survey.jl copied to clipboard
JK1 as option in replicate weighting
Based on discussion about JKn taking some time to think and implement, and the package currently having a JK1 working for the designs it is meant to work for (simple, strat and basic cluster), we should integrate the current jackknife
as what it is, instead of removing it entirely.
So have a jk1_weights
kind of function which works when certain conditions of the SurveyDesign
are met? Thoughts?
#137 can be looked into thoroughly at later stage
I think we should directly aim for JKn. JK1 doesn't work on stratified sample.
R default is Jk1 for non stratified, JKn if strata. must be for some reasons.