cluster.resample {cfdecomp}

Cluster Resampling: resampling long format longitudinal or otherwise clustered data

Description

Cluster Resampling: resampling long format longitudinal or otherwise clustered data

Usage

cluster.resample(data, cluster.name, size = NA, newID = FALSE)

Arguments

data

a data frame containing the variables in the model.

cluster.name

the name (as a character) of the column containing the cluster identifiers.

size

a non-negative integer giving the number of items to choose, i.e. the number of clusters to resample. If not specified, takes the original data cluster size.

newID

if set to TRUE, gives each replicant a new ID, rather than having the old ID appear multiple times. This might be needed if you work with e.g. cluster fixed effects

Value

returns a new dataframe with resampled clusters

Examples

table(cfd.example.data$id)
# every ID appears 5 times
cfd.example.sample <- cluster.resample(cfd.example.data, cluster.name='id')
table(cfd.example.sample$id)
# some ID's now don't appear, and some appear more times (multiples of 5)
# the important part is that if a person (id) is resampled, all their rows of data are taken
# i.e. the function resamples clusters, rather than rows.
# this has produced 1 resample, so generally this function would be used
# inside a loop where it is used multiple times
# we use this function inside our decomposition functions when cluster.sample is
# set to TRUE in those functions.