Cluster sampling

Selects whole clusters at random and returns all of their rows. The row count follows from which clusters were picked, so there is no n; use design_multistage() when you need to control it.

Usage

design_cluster(clusters, n_clusters, balanced = FALSE, na_rm = FALSE)

Arguments

clusters

Column naming each row's cluster.

n_clusters

Number of clusters to select.

balanced

Take an equal number of rows from each selected cluster, equal to the smallest selected cluster's size.

na_rm

Drop rows whose cluster label is NA instead of raising an error. When FALSE, missing labels are never treated as a cluster of their own.

Value

A design object, for use with draw().

See also

draw()

Other designs: design_bootstrap(), design_certainty(), design_multistage(), design_reservoir(), design_simple(), design_spatial(), design_spread(), design_stratified(), design_systematic(), design_temporal(), design_weighted()

Examples

df <- data.frame(id = 1:100, site = rep(paste0("s", 1:10), each = 10))
unique(draw(df, design_cluster("site", n_clusters = 3), seed = 1)$site)
#> [1] "s4" "s7" "s9"