pairwise_ordcat_test: Pairwise comparison for ordinal categories

In wrappedtools: Useful Wrappers Around Commonly Used Functions

Pairwise comparison for ordinal categories

Description

pairwise_ordcat_test calculates pairwise comparisons for ordinal categories between all group levels with corrections for multiple testing.

Usage

pairwise_ordcat_test(
  dep_var,
  indep_var,
  adjmethod = "fdr",
  plevel = 0.05,
  symbols = letters[-1],
  ref = FALSE,
  cmh = TRUE
)

Arguments

dep_var	dependent variable, containing the data
indep_var	independent variable, should be factor
adjmethod	method for adjusting p values (see p.adjust)
plevel	threshold for significance
symbols	predefined as b,c, d...; provides footnotes to mark group differences, e.g. b means different from group 2
ref	is the 1st subgroup the reference (like in Dunnett test)
cmh	Should Cochran-Mantel-Haenszel test (coin::cmh_test) be used for testing? If false, the linear-by-linear association test (coin::lbl_test) is applied.

Value

A list with elements "methods" (character), "p.value" (matrix), "plevel" (numeric), and "sign_colwise" (vector of length number of levels - 1)

Examples

# All pairwise comparisons
mtcars2 <- dplyr::mutate(mtcars, cyl = factor(cyl, ordered = TRUE))
pairwise_ordcat_test(dep_var = mtcars2$cyl, indep_var = mtcars2$gear)
# Only comparison against reference gear=3
pairwise_ordcat_test(dep_var = mtcars2$cyl, indep_var = mtcars2$gear, ref = TRUE)

wrappedtools documentation built on Oct. 10, 2024, 5:07 p.m.