Calculate number of level 1 and level 2 variables — count

Given a control formula, determine which variables are level 1 and which are level 2. Level 1 variables vary within clusters, level 2 do not. If formula and data are provided, will first convert to canonical form. If just_count is TRUE, will return a named vector with counts of level 1 and level 2 variables.

Usage

count_covariates(
  data,
  control_formula,
  cluster_formula = NULL,
  just_count = FALSE,
  pure = TRUE
)

Arguments

data: Data frame with outcome, treatment, clusterID, blockID, and covariates.
control_formula: Formula with covariates only (no outcome or treatment).
cluster_formula: Formula for cluster ID, of form ~ CID or ~ CID | BID. In the latter case, will make interaction to ensure all clusters have unique ID. If null, assumes a clusterID and blockID column exist in the data.
just_count: If TRUE, return just counts of level 1 and level 2 variables.
pure: TRUE means count level 2 covariates ONLY if they are ONLY level 2. FALSE means covariates that are both level 1 and level 2 get counted twice, once for level 1 and once for level 2.

Value

If just_count is FALSE (default), a tibble with columns:

variable: Variable name.
level: Either "Level-1" or "Level-2".

If just_count is TRUE, a named vector with counts of level 1 and level 2 variables.