-
Fix issue #32:
stack_replicate_designs()
would only accept designs with types known to the 'survey' package. Fixed to allow other design types such as the random-groups jackknife. -
Added new helper function
shuffle_replicates()
to randomize the order of the columns of replicate weights. This is useful, for example, when replicates are created independently in each stratum and then combined. -
Added new helper function
subsample_replicates()
to retain a random subset of replicates and accordingly increase the scale factor used for variance estimation. -
Added new helper function
add_inactive_replicates()
which allows the user to add "inactive" replicates to a design object, so that the matrix of replicate weights has the desired number of columns. -
The function
as_fays_gen_rep_design()
now has a default value ofmse = TRUE
. Settingmse = FALSE
will produce a warning message, since Fay's generalized replication method can sometimes produce large underestimates of variance whenmse = FALSE
. -
The function
as_random_group_jackknife_design()
now returns an object with classtbl_svy
if the input was also an object with classtbl_svy
. -
Generalized replication based on the Yates-Grundy estimator is now much faster due to code refactoring.