blockclustr: Constructs blocked clusters for blocked and clustered random...

In erossiter/blockclustr: Performs block cluster randomization when units also need to be clustered

Constructs blocked clusters for blocked and clustered random assignment

Description

Constructs blocked clusters for blocked and clustered random assignment

Usage

blockclustr(
  data,
  id_var,
  block_vars,
  n_tr,
  n_units_per_clust,
  constraint_var = NULL,
  constraint_list = NULL,
  algorithm = "optGreedy",
  distance = "mahalanobis"
)

Arguments

data	A dataframe with units in rows and variables in columns.
id_var	A string specifying which column of data is the units' unique identifier
block_vars	A string or vector of strings specifying which columns of data are the (continuous) blocking covariates.
n_tr	An integer specifying the number of experimental conditions. This is also the number of clusters per block.
n_units_per_clust	An integer specifying the number of units that should be assigned to each cluster. It will not be used if constraint_list is specified.
constraint_var	An optional string specifying which column of data is the variable that cluster composition is constrained by. Must also specify constraint_list.
constraint_list	An optional list of vectors specifying all combinations of constraint_var values that clusters can take. Must also specify constraint_var.
algorithm	A string specifying the algorithm used to construct blockettes. See blockTools::block for more options and information.
distance	A string specifying the distance calculation used when constructing blockettes. See blockTools::block for more options and information.

Value

Returns a list of three elements:

• design a dataframe containing information about units included in the experimental design

  • id_var the supplied unique identifier

  • block_vars the supplied blocking covariates

  • constraint_var the supplied constraint variable, if applicable

  • block_id an indicator for which block the unit belongs to

  • block_id an indicator for which cluster the unit belongs to

  • block_id a treatment indicator, corresponding with the specified number of conditions, n_tr

• raw_blockettes is the output of the blockTools::block function, used to construct blockettes

• unused_units a dataframe containing information about the units in data not included in the experimental design

erossiter/blockclustr documentation built on March 19, 2022, 4:48 a.m.