Power of one-way ANOVA given Cohen's f

Usage

mrm_anova_power(k_groups, n_per_group, effect_size_f, alpha = 0.05)

Arguments

k_groups: Number of groups.
n_per_group: Per-group sample size.
effect_size_f: Cohen's f.
alpha: Type-I error (default 0.05).

Value

Named list with k_groups, n_per_group, N_total, effect_size_f, alpha, df1, df2, noncentrality, F_critical, power, interpretation.

Examples

# Power to detect a medium effect (Cohen's f = 0.25) with 4 groups
# of 30 each at alpha = 0.05:
res <- mrm_anova_power(
  k_groups = 4, n_per_group = 30,
  effect_size_f = 0.25, alpha = 0.05
)
res$power
#> [1] 0.6065
res$F_critical
#> [1] 2.6828

# Sample-size sensitivity: what power do I get with smaller groups?
sapply(c(10, 20, 30, 50, 100), function(n) {
  mrm_anova_power(
    k_groups = 3, n_per_group = n,
    effect_size_f = 0.25
  )$power
})
#> [1] 0.1951 0.3744 0.5396 0.7796 0.9781