Null distribution for minimum p-values

Computes the distribution of the minimum p-value, particularly Monte Carlo samples from the combined null distribution F in Theorem 1 of the paper.

Usage

comb_null_dist_cre(
  n,
  m,
  methods.list,
  Z.perm = NULL,
  nperm = 10^4,
  stat.null.mult = NULL
)

Arguments

n

Total number of units.

m

Number of treated units.

methods.list

A list of method specifications.

Z.perm

Optional permutation matrix.

nperm

Number of permutations for null distribution.

stat.null.mult

Optional pre-computed null distribution matrix.

Value

A numeric vector of minimum p-values under the null. Its length is the number of draws actually available: ncol(Z.perm) when a permutation matrix is supplied, ncol(stat.null.mult) when a null distribution is, and nperm otherwise.

Examples

set.seed(1)
methods.list <- list(list(name = "Wilcoxon", scale = FALSE),
                     list(name = "Stephenson", s = 3, scale = FALSE))
cnd <- comb_null_dist_cre(n = 20, m = 12, methods.list = methods.list,
                          nperm = 500)
quantile(cnd, c(0.05, 0.10))
#>     5%    10% 
#> 0.0459 0.0874