Holm step-down FWER procedure

Thin wrapper over stats::p.adjust(method = "holm"); uniformly more powerful than Bonferroni.

Usage

holm(p_values, alpha = 0.05, labels = NULL)

Arguments

p_values: Numeric vector of raw p-values.
alpha: Significance level.
labels: Optional character vector of test labels.

Value

An object of class "morie_multiple_testing_result".

Examples

set.seed(1)
p <- c(runif(30), runif(5, 0, 0.005))

res <- holm(p)
res$n_rejected
#> [1] 1
head(res$adjusted)
#> [1] 1 1 1 1 1 1

# Holm is uniformly more powerful than Bonferroni (rejects at least as
# many), while still controlling the family-wise error rate:
c(holm = holm(p)$n_rejected, bonferroni = bonferroni(p)$n_rejected)
#>       holm bonferroni 
#>          1          1 

# alpha + labels as usual.
holm(c(0.001, 0.01, 0.04), alpha = 0.05,
     labels = c("A", "B", "C"))$rejected
#> [1] TRUE TRUE TRUE