tests/testthat/test_pval.R

In CATTexact: Computation of the p-Value for the Exact Conditional Cochran-Armitage Trend Test

library(CATTexact)
context("P-values")


test_that("Comparing p-values with reference", {
  expect_equal(round(catt_exact(c(0,1,2), c(5,5,5), c(3,3,3))$exact.pvalue,3),0.622)
  expect_equal(round(catt_exact(c(0,1,2), c(5,5,5), c(3,3,3))$asymptotic.pvalue,3),0.5)
  expect_equal(round(catt_asy(c(0,1,2), c(5,5,5), c(3,3,3))$asymptotic.pvalue,3),0.5)
  expect_equal(round(catt_exact(c(0,10,35), c(48,50,52), c(3,13,16))$exact.pvalue,3),0.005)
  expect_equal(round(catt_exact(c(0,10,35), c(48,50,52), c(3,13,16))$asymptotic.pvalue,3),0.004)
  expect_equal(round(catt_asy(c(0,10,35), c(48,50,52), c(3,13,16))$asymptotic.pvalue,3),0.004)
})

test_that("Output is correct when dose responses are noninteger (known bug in earlier versions)", {
  expect_equal(round(catt_exact(c(0,85.5,285.2,1077.4), c(51,51,51,51), c(0,2,1,1))$exact.pvalue,3),0.418)
  expect_equal(round(catt_exact(c(0,85.5,285.2,1077.4), c(51,51,51,51), c(0,2,1,1))$asymptotic.pvalue,3),0.460)
  expect_equal(round(catt_asy(c(0,85.5,285.2,1077.4), c(51,51,51,51), c(0,2,1,1))$asymptotic.pvalue,3),0.460)
  expect_equal(round(catt_exact(c(0,0.001,0.1,1,10), c(50,50,50,50,50), c(6,8,10,11,10))$exact.pvalue,3),0.283)
})