test-bf_ttest.R

# bayes factor (independent samples t-test) ----------------------

testthat::test_that(
  desc = "bayes factor (independent samples t-test)",
  code = {
    testthat::skip_if(getRversion() < "3.6")

    # from Bayes Factor
    df <- suppressMessages(bf_extractor(
      BayesFactor::ttestBF(
        formula = len ~ supp,
        data = as.data.frame(ToothGrowth),
        rscale = 0.99,
        paired = FALSE
      )
    ))

    # extracting results from where this function is implemented
    set.seed(123)
    df_results <-
      bf_ttest(
        data = ToothGrowth,
        x = supp,
        y = "len",
        paired = FALSE,
        bf.prior = 0.99
      )

    # check bayes factor values
    testthat::expect_is(df, "tbl_df")
    testthat::expect_equal(df$log_e_bf10, -0.001119132, tolerance = 0.001)

    # checking if two usages of the function are producing the same results
    testthat::expect_equal(df$bf10, df_results$bf10, tolerance = 0.001)
  }
)

# Bayes factor (paired t-test) ---------------------------------------------

testthat::test_that(
  desc = "bayes factor (paired t-test)",
  code = {
    testthat::skip_if(getRversion() < "3.6")

    # data
    dat <- tidyr::spread(bugs_long, condition, desire) %>%
      dplyr::filter(.data = ., !is.na(HDLF), !is.na(HDHF))

    # BF output
    set.seed(123)
    df <- suppressMessages(bf_extractor(
      BayesFactor::ttestBF(
        x = dat$HDLF,
        y = dat$HDHF,
        rscale = 0.8,
        paired = TRUE
      )
    ))

    # creating a tidy dataframe
    dat_tidy <- dplyr::filter(bugs_long, condition %in% c("HDLF", "HDHF"))

    # extracting results from where this function is implemented
    set.seed(123)
    df_results <-
      bf_ttest(
        data = dat_tidy,
        x = "condition",
        y = desire,
        paired = TRUE,
        bf.prior = 0.8
      )

    # check bayes factor values
    testthat::expect_equal(df$bf10, 40.36079, tolerance = 0.001)
    testthat::expect_equal(df$log_e_bf10, 3.697859, tolerance = 0.001)

    # checking if two usages of the function are producing the same results
    testthat::expect_equal(df$bf10, df_results$bf10, tolerance = 0.001)
  }
)

# bayes factor (one sample t-test) ----------------------

testthat::test_that(
  desc = "bayes factor (one sample t-test)",
  code = {
    testthat::skip_if(getRversion() < "3.6")

    # creating a dataframe
    set.seed(123)
    df <-
      suppressMessages(bf_extractor(
        BayesFactor::ttestBF(
          x = iris$Petal.Length,
          mu = 5.5,
          rscale = 0.99
        )
      ))

    # extracting results from where this function is implemented
    set.seed(123)
    df_results <-
      bf_ttest(
        data = iris,
        x = Petal.Length,
        y = NULL,
        test.value = 5.5,
        bf.prior = 0.99
      )

    # check Bayes factor values
    testthat::expect_equal(df$bf10, 5.958171e+20, tolerance = 0.001)
    testthat::expect_equal(df$log_e_bf10, 47.83647, tolerance = 0.001)

    # checking if two usages of the function are producing the same results
    testthat::expect_equal(df$bf10, df_results$bf10, tolerance = 0.001)

    # TO DO: wait for `easystats` to be updated
    # # extracting subtitle (without NA)
    # set.seed(123)
    # subtitle <-
    #   bf_ttest(
    #     data = iris,
    #     x = "Petal.Length",
    #     y = NULL,
    #     test.value = 5.5,
    #     bf.prior = 0.99,
    #     output = "subtitle",
    #     centrality = "mean",
    #     conf.level = 0.90
    #   )
    #
    # testthat::expect_identical(
    #   subtitle,
    #   ggplot2::expr(
    #     atop(
    #       displaystyle(NULL),
    #       expr = paste(
    #         "log"["e"],
    #         "(BF"["10"],
    #         ") = ",
    #         "47.84",
    #         ", ",
    #         widehat(italic(d))["mean"]^
    #           "posterior",
    #         " = ",
    #         "1.76",
    #         ", CI"["90%"]^"HDI",
    #         " [",
    #         "1.52",
    #         ", ",
    #         "1.99",
    #         "]",
    #         ", ",
    #         italic("r")["Cauchy"]^"JZS",
    #         " = ",
    #         "0.99"
    #       )
    #     )
    #   )
    # )
    #
    # # extracting subtitle (with NA)
    # set.seed(123)
    # subtitle2 <-
    #   bf_ttest(
    #     data = ggplot2::msleep,
    #     x = brainwt,
    #     y = NULL,
    #     test.value = 0.25,
    #     bf.prior = 0.9,
    #     output = "subtitle",
    #     conf.method = "eti"
    #   )
    #
    # testthat::expect_identical(
    #   subtitle2,
    #   ggplot2::expr(atop(
    #     displaystyle(NULL),
    #     expr = paste(
    #       "log"["e"],
    #       "(BF"["10"],
    #       ") = ",
    #       "-2.13",
    #       ", ",
    #       widehat(italic(d))["median"]^
    #         "posterior",
    #       " = ",
    #       "-0.02",
    #       ", CI"["95%"]^"ETI",
    #       " [",
    #       "-0.27",
    #       ", ",
    #       "0.23",
    #       "]",
    #       ", ",
    #       italic("r")["Cauchy"]^
    #         "JZS",
    #       " = ",
    #       "0.90"
    #     )
    #   ))
    # )
  }
)