- blocks
A vector of length N indicating which block each unit belongs to. Every unit in a cluster must belong to the same block. (required)
- clusters
A vector of length N indicating which cluster each unit belongs to. (required)
- prob
Use for a two-arm design in which either floor(N_clusters_block*prob) or ceiling(N_clusters_block*prob) clusters are assigned to treatment within each block. Which of the two is used is itself random: the ceiling is drawn with probability equal to the fractional part of N_clusters_block*prob and the floor otherwise, which makes each cluster's probability of assignment exactly prob. Must be a real number between 0 and 1 inclusive. (optional)
- prob_unit
Use for a two-arm design. Must be of length N. tapply(prob_unit, blocks, unique) will be passed to block_prob, so it must be constant within each block. (optional)
- prob_each
Use for a multi-arm design. A numeric vector giving the probability of assignment to each condition. All entries must be between 0 and 1 inclusive and must sum to 1. Because of integer rounding, the exact number of clusters assigned to each condition may differ slightly from assignment to assignment, but the overall probability of assignment is exactly prob_each. (optional)
- m
Use for a two-arm design in which the scalar m gives the fixed number of clusters assigned to treatment within every block. This count does not vary across blocks. (optional)
- m_unit
Use for a two-arm design. Must be of length N. tapply(m_unit, blocks, unique) will be passed to block_m, so it must be constant within each block. (optional)
- block_m
Use for a two-arm design in which block_m gives the number of clusters to assign to treatment within each block. Must be a numeric vector as long as the number of blocks, in the same order as sort(unique(blocks)). (optional)
- block_m_each
Use for a multi-arm design in which block_m_each gives the number of clusters assigned to each condition within each block. Must be a matrix with one row per block and one column per treatment arm. Rows respect the ordering of blocks by sort(unique(blocks)); columns should be in the order of conditions, if specified. (optional)
- block_prob
Use for a two-arm design in which the probability of assignment to treatment varies across blocks. Must be in the same order as sort(unique(blocks)). Differs from prob in that the probability of assignment can vary across blocks. (optional)
- block_prob_each
Use for a multi-arm design in which assignment probabilities vary across blocks. Must be a matrix with one row per block and one column per treatment arm; each row must sum to 1. Rows respect the ordering of sort(unique(blocks)). Use only if the probabilities of assignment should vary by block, otherwise use prob_each. (optional)
- num_arms
The number of treatment arms. If unspecified, determined from the other arguments. (optional)
- conditions
A character vector giving the names of the treatment groups. If unspecified, the treatment groups will be named 0 (for control) and 1 (for treatment) in a two-arm trial and T1, T2, T3, in a multi-arm trial. A two-group design in which num_arms is set to 2 will use condition names T1 and T2. (optional)
- check_inputs
Logical. Whether to verify before assigning that the arguments are internally consistent: that clusters nest within blocks, that counts sum to the number of clusters in each block, that probabilities lie between 0 and 1 and sum to 1, and so on. Defaults to TRUE. FALSE skips the checking only: num_arms and conditions are still derived from the other arguments, so the same call draws the same assignment either way. What goes is the verification, and an impossible design is then no longer refused. block_m larger than a block, for instance, quietly treats the whole block. Declaring the design once with declare_ra() and drawing from it with conduct_ra() is the usual way to avoid re-checking the same arguments in a simulation. (optional)