Extreme p-values

readme.md

extreme p_values.R

library(tidyverse)
library(ggdist)



data = expand.grid(
  effect_size = c(0, 0.2, 0.4, 0.6, 0.8),
  sample_size = 1000,
  rep = 1:10000
) %>% 
  rowwise() %>% 
  mutate(p = t.test(rnorm(sample_size/2), rnorm(sample_size/2, effect_size))$p.value)

ggplot(data) +
  aes(x = p, fill = factor(effect_size), group = effect_size, y = factor(effect_size)) +
  ggdist::stat_histinterval(linetype = 0, point_alpha = 0, normalize = "groups") +
  scale_x_log10() +
  theme_minimal(14) + theme(legend.position = "top", legend.justification = "left") +
  labs(x = "p-values", y = NULL, fill = "Cohen's d",
    title = "Simulating p-values for sample size 1000")



data_ss = expand.grid(
  effect_size = 0.5,
  sample_size = c(50, 100, 500, 1000, 5000),
  rep = 1:10000
) %>% 
  rowwise() %>% 
  mutate(p = t.test(rnorm(sample_size/2), rnorm(sample_size/2, effect_size))$p.value)

ggplot(data_ss) +
  aes(x = p, fill = factor(sample_size), group = sample_size, y = factor(sample_size)) +
  ggdist::stat_histinterval(linetype = 0, point_alpha = 0, normalize = "groups") +
  scale_x_log10() +
  theme_minimal(14) + theme(legend.position = "top", legend.justification = "left") +
  labs(x = "p-values", y = NULL, fill = "Sample size",
    title = "Simulating p-values for Cohen's d = 0.5")