Bootstrap confidence intervals and p values

Usage

bootstrap_estimates(
  vals,
  grps,
  reference_arm = sort(unique(grps))[[1]],
  na_replacement = NULL,
  n_samples = 2000,
  alpha = 0.05
)

vals: numeric vector
grps: character vector, allowed values are "ctrl" and "actv" and used to slice x into two vectors with values to be compared.
reference_arm: character vector of length 1, name of the reference arm in grps
na_replacement: scalar (or NULL), if NULL missing values will be removed. Otherwise, missing values will be replaced by the value of na_replacement
n_samples: single integer, number of bootstrap samples
alpha: numeric, 1.0 - confidence_level

A four-element vector with point estimate, p value and confidence internal