Create generalized replication replicate weights

Generates Fay's generalized replication weights via svrep::as_fays_gen_rep_design(). Produces deterministic replicate weights (no randomness). Requires a survey_taylor design.

Usage

create_gen_rep_weights(
  data,
  ...,
  variance_estimator = "SD2",
  max_replicates = Inf,
  balanced = TRUE,
  aux_var_names = NULL,
  mse = TRUE,
  seed = NULL
)

Arguments

data

A survey_taylor design.

...

Must be empty.

variance_estimator

character(1). Target variance estimator. Same options as create_gen_boot_weights(). Default "SD2".

max_replicates

numeric(1), default Inf. Maximum number of replicates; Inf uses the natural count.

balanced

logical(1), default TRUE. Equal contribution of replicates to variance estimates.

aux_var_names

<tidy-select> or NULL. Required for "Deville-Tille".

mse

logical(1), default TRUE.

seed

integer(1) or NULL. RNG seed for reproducibility.

Value

A survey_replicate with generalized replication weights.

See also

Other replicate-weights: as_taylor_design(), create_bootstrap_weights(), create_brr_weights(), create_gen_boot_weights(), create_jackknife_weights(), create_replicate_weights(), create_sdr_weights()