permutest (version 1.0.0)

permutation_test_ci: Construct confidence interval by inverting permutation tests

Description

This function constructs a confidence interval by inverting permutation tests and applying the method in Glazer and Stark, 2024.

Usage

permutation_test_ci(
  df,
  group_col,
  outcome_col,
  strata_col = NULL,
  test_stat = "diff_in_means",
  perm_func = permute_group,
  upper_bracket = NULL,
  lower_bracket = NULL,
  cl = 0.95,
  e = 0.1,
  reps = 10000,
  perm_set = NULL,
  seed = 42
)

Value

A list containing the permutation test p-value, and the test statistic distribution if applicable

Arguments

df

A data frame

group_col

The name of the column in df that corresponds to the group label

outcome_col

The name of the column in df that corresponds to the outcome variable

strata_col

The name of the column in df that corresponds to the strata

test_stat

Test statistic function

perm_func

Function to permute group

upper_bracket

Array with 2 values that bracket upper confidence bound

lower_bracket

Array with 2 values that bracket lower confidence bound

cl

Confidence level, default 0.95

e

Maximum distance from true confidence bound value

reps

Number of iterations to use when calculating permutation p-value

perm_set

Matrix of group assignments to use instead of reps iterations of perm_func

seed

An integer seed value

Examples

x <- c(35.3, 35.9, 37.2, 33.0, 31.9, 33.7, 36.0, 35.0, 33.3, 33.6, 37.9, 35.6, 29.0, 33.7, 35.7)
y <- c(32.5, 34.0, 34.4, 31.8, 35.0, 34.6, 33.5, 33.6, 31.5, 33.8, 34.6)
df <- data.frame(outcome = c(x, y), group = c(rep(1, length(x)), rep(0, length(y))))
permutation_test_ci(df = df, group_col = "group", outcome_col = "outcome", strata_col = NULL,
                    test_stat = "diff_in_means", perm_func = permute_group,
                    upper_bracket = NULL, lower_bracket = NULL,
                    cl = 0.95, e = 0.01, reps = 10^3, seed = 42)