tests/testthat/testclustered_se.R

In mattdwebb/wildclusterboot: Wild Cluster Bootstrap

library(testthat)
library(wildclusterboot)

test_that('clustered_se calculates proper values',{

  #Set out X and Y vectors
  X <- c(1.9157903, 1.9157903, 1.4053975, 1.4053975, 0.3096093, 0.3096093, -0.2007836, -0.2007836)
  Y <- c(5.3051725, 3.2240347, 0.3796652, 1.4897371, 2.4207392, 3.4025150, 0.6618861, 1.4311579)

  #Set out clusterby vector
  G <- c(1, 1, 2, 2, 1, 1, 2, 2)

  data <- data.frame(X = X, Y = Y, G = G)

  #Result of what se vector should be
  test_se <- c('(Intercept)' = 0.7938156, 'X' = 0.4473510)

  #Generate model and test if result is expected value and size
  model <- lm(data = data, formula = Y ~ X)
  se <- clustered_se(data = data, model = model, clusterby = 'G')
  expect_equal(se, test_se, tolerance = 0.0001)
  expect_equal(length(se),2)

  #Create second X value and check that resulting vector correct size and names
  X2 <- rnorm(length(X))
  data2 <- cbind(data, X2)
  model <- lm(Y ~ X + X2)
  se <- clustered_se(data = data2, model = model, clusterby = 'G')
  expect_equal(length(se), 3)
  expect_equal(names(se), c('(Intercept)', 'X', 'X2'))

  #Make sure throws correct error on model not lm class
  expect_error(clustered_se(model = '', clusterby = 'G'),
               'Model variable must be lm class')

})

test_that('clustered_se calculates proper values for multiway',{

  #Set out X and Y vectors
  X <- c(1.9157903, 1.9157903, 1.4053975, 1.4053975, 0.3096093, 0.3096093, -0.2007836, -0.2007836)
  Y <- c(5.3051725, 3.2240347, 0.3796652, 1.4897371, 2.4207392, 3.4025150, 0.6618861, 1.4311579)

  #Set out clusterby vector
  G <- c(1, 1, 2, 2, 1, 1, 2, 2)
  H <- c(1, 2, 1, 2, 1, 2, 1, 2)

  data <- data.frame(X = X, Y = Y, G = G, H = H)
  clusterby = ~ G + H

  #Result of what se vector should be
  test_se <- c('(Intercept)' = 0.8054447, 'X' = 0.5150004)

  #Generate model and test if result is expected value and size
  model <- lm(data = data, formula = Y ~ X)
  se <- clustered_se(data = data, model = model, clusterby = clusterby)
  expect_equal(se, test_se, tolerance = 0.0001)
  expect_equal(length(se), 2)

  #Set out X and Y vectors
  data <- read.csv('test_files/multiway.csv')

  clusterby <- ~ G + H + F

  #Result of what se vector should be
  test_se <- c('(Intercept)' = 1.0919700, 'X' = 0.1106969)

  #Generate model and test if result is expected value and size
  model <- lm(data = data, formula = Y ~ X)
  se <- clustered_se(data = data, model = model, clusterby = clusterby)
  expect_equal(se, test_se, tolerance = 0.0001)
  expect_equal(length(se), 2)

  #Test of spectral decomposition correction works
  test_se <- c('(Intercept)' = 0.0109916, 'W' = 0.0032928)

  data <- read.csv('test_files/multiway_problems.csv')
  model <- lm(data = data, formula = Y ~ W)
  clusterby <- ~ G + H + F
  se <- clustered_se(data = data, model = model, clusterby = clusterby)
  expect_equal(se, test_se, tolerance = 0.0001)

})