Prepare CRAN submission

.Rbuildignore

@@ -24,3 +24,4 @@
 ^CRAN-SUBMISSION$
 \.code-workspace$
 ^\.lintr$
+^tests/testthat/_snaps/.

DESCRIPTION

@@ -1,7 +1,7 @@
 Type: Package
 Package: modelbased
 Title: Estimation of Model-Based Predictions, Contrasts and Means
-Version: 0.8.9.42
+Version: 0.9.0
 Authors@R:
     c(person(given = "Dominique",
              family = "Makowski",
@@ -89,4 +89,3 @@ Config/testthat/parallel: true
 Roxygen: list(markdown = TRUE)
 Config/Needs/check: stan-dev/cmdstanr
 Config/Needs/website: easystats/easystatstemplate
-Remotes: vincentarelbundock/marginaleffects

R/estimate_contrasts.R

@@ -52,7 +52,7 @@
 #' # By default: selects first factor
 #' estimate_contrasts(model)
 #'
-#' # Can also run contrasts between points of numeric
+#' # Can also run contrasts between points of numeric, stratified by "Species"
 #' estimate_contrasts(model, contrast = "Petal.Width", by = "Species", length = 4)
 #'
 #' # Or both

R/estimate_means.R

@@ -49,21 +49,21 @@
 #'   which computes a kind of "weighted average" for the values at which these
 #'   terms are hold constant. These predictions are a good representation of the
 #'   sample, because all possible values and levels of the non-focal predictors
-#'   are taken into account. It answers the question, "What is the predicted
-#'   value for an 'average' observation in *my data*?". It refers to randomly
-#'   picking a subject of your sample and the result you get on average. This
-#'   approach is the one taken by default in the `emmeans` package.
+#'   are considered. It answers the question, "What is the predicted value for
+#'   an 'average' observation in *my data*?". It refers to randomly picking a
+#'   subject of your sample and the result you get on average. This approach is
+#'   the one taken by default in the `emmeans` package.
 #' - `"population"`: Non-focal predictors are marginalized over the observations
 #'   in the sample, where the sample is replicated multiple times to produce
 #'   "counterfactuals" and then takes the average of these predicted values
 #'   (aggregated/grouped by the focal terms). It can be considered as
 #'   extrapolation to a hypothetical target population. Counterfactual
 #'   predictions are useful, insofar as the results can also be transferred to
 #'   other contexts (Dickerman and Hernan, 2020). It answers the question, "What
-#'   is the predicted for the 'average' observation in *the broader target
-#'   population*?". It does not only refer to the actual data in your observed
-#'   sample, but also "what would be if" we had more data, or if we had data
-#'   from a different sample.
+#'   is the predicted response value for the 'average' observation in *the
+#'   broader target population*?". It does not only refer to the actual data in
+#'   your observed sample, but also "what would be if" we had more data, or if
+#'   we had data from a different sample.
 #'
 #' In other words, the distinction between marginalization types resides in whether
 #' the prediction are made for:

tests/testthat/test-attributes_estimatefun.R

@@ -1,5 +1,3 @@
-## TODO: skip for marginaleffects 0.24.0?
-
 skip_if_not_installed("emmeans")
 skip_if_not_installed("marginaleffects")
 
@@ -55,6 +53,7 @@ test_that("attributes_means, contrasts", {
 
 
 test_that("attributes_means, slopes", {
+  skip_if_not_installed("marginaleffects", minimum_version = "0.24.1")
   data(iris)
   model <- lm(Sepal.Length ~ Species + Sepal.Width, data = iris)
 

tests/testthat/test-brms.R

@@ -1,3 +1,4 @@
+skip_on_cran()
 skip_on_os("windows")
 skip_if_not_installed("brms")
 skip_if_not_installed("BH")

tests/testthat/test-estimate_contrasts.R

@@ -22,7 +22,7 @@
   estim <- suppressMessages(estimate_contrasts(model, by = "Species=c('versicolor', 'virginica')", backend = "emmeans"))
   expect_identical(dim(estim), c(1L, 9L))
 
   estim <- suppressMessages(estimate_contrasts(model, contrast = "Species=c('versicolor', 'virginica')", backend = "marginaleffects"))
   expect_identical(dim(estim), c(1L, 9L))
 
   # Two factors
@@ -40,7 +40,7 @@
   expect_identical(dim(estim), c(3L, 10L))
   estim <- suppressMessages(estimate_contrasts(model, contrast = "Species", backend = "marginaleffects"))
   expect_identical(dim(estim), c(3L, 9L))
   estim <- suppressMessages(estimate_contrasts(model, contrast = "Species", by = "fac='A'", backend = "marginaleffects"))
   expect_identical(dim(estim), c(3L, 9L))
 
   # One factor and one continuous
@@ -83,7 +83,7 @@
   ## FIXME: doesn't work right nw
   # estim <- suppressMessages(estimate_contrasts(model, by = "all", backend = "marginaleffects"))
   # expect_identical(dim(estim), c(12L, 11L))
   estim <- suppressMessages(estimate_contrasts(model, contrast = c("vs", "am"), by = "gear='5'", backend = "marginaleffects"))
   expect_identical(dim(estim), c(6L, 9L))
   expect_named(estim, c("Level1", "Level2", "Difference", "SE", "CI_low", "CI_high", "t", "df", "p"))
   expect_equal(estim$Difference, c(-6.98333, -11.275, -18.25833, -4.29167, -11.275, -6.98333), tolerance = 1e-4)
@@ -107,11 +107,11 @@
 
   model <- lm(Petal.Width ~ factor1 * factor2 * factor3, data = dat)
 
   estim <- suppressMessages(estimate_contrasts(model, contrast = c("factor1", "factor2", "factor3"), by = "all", backend = "emmeans"))
   expect_identical(dim(estim), c(28L, 9L))
   estim <- suppressMessages(estimate_contrasts(model, contrast = c("factor1", "factor2"), by = "factor3='F'", backend = "emmeans"))
   expect_identical(dim(estim), c(6L, 10L))
   estim <- suppressMessages(estimate_contrasts(model, contrast = c("factor1", "factor2"), by = "factor3", backend = "emmeans"))
   expect_identical(dim(estim), c(12L, 10L))
 
 
@@ -155,6 +155,7 @@
 
 
 test_that("estimate_contrasts - Bayesian", {
+  skip_on_cran()
   skip_if_not_installed("logspline")
   skip_if_not_installed("rstanarm")
   skip_if_not_installed("lme4")
@@ -178,7 +179,7 @@
   )
   estim <- suppressMessages(estimate_contrasts(model, contrast = "all", backend = "emmeans"))
   expect_identical(dim(estim), c(15L, 7L))
   estim <- suppressMessages(estimate_contrasts(model, by = c("Species", "Petal.Length_factor='A'"), backend = "emmeans"))
   expect_identical(dim(estim), c(3L, 8L))
   estim <- estimate_contrasts(model, contrast = c("Species", "Petal.Length_factor"), backend = "marginaleffects")
   expect_identical(dim(estim), c(15L, 10L))
@@ -234,7 +235,7 @@
 
   estim <- suppressWarnings(suppressMessages(estimate_contrasts(model, test = "bf", backend = "emmeans")))
   expect_identical(dim(estim), c(3L, 6L))
   estim <- suppressWarnings(suppressMessages(estimate_contrasts(model, predict = "link", test = "bf", backend = "emmeans")))
   expect_identical(dim(estim), c(3L, 6L))
 })
 
@@ -254,6 +255,7 @@
 
 
 test_that("estimate_contrasts - dfs", {
+  skip_on_cran()
   skip_if_not_installed("lme4")
   skip_if_not_installed("pbkrtest")
   skip_if_not_installed("lmerTest")
@@ -263,8 +265,8 @@
   data$Petal.Length_factor <- ifelse(data$Petal.Length < 4.2, "A", "B")
   model <- lme4::lmer(Sepal.Width ~ Species + (1 | Petal.Length_factor), data = data)
 
   estim1 <- suppressMessages(estimate_contrasts(model, lmer.df = "satterthwaite", p_adjust = "holm", backend = "emmeans"))
   estim2 <- suppressMessages(estimate_contrasts(model, lmer.df = "kenward-roger", p_adjust = "holm", backend = "emmeans"))
 
   expect_true(all(estim1$CI_low != estim2$CI_low))
   expect_equal(estim1$CI_low, c(-2.43, -2.25692, -2.89384), tolerance = 1e-4)
@@ -280,7 +282,7 @@
 
 
 test_that("estimate_contrasts - marginaleffects", {
-  ## TODO: skip for marginaleffects 0.24.0?
+  # ignore failing snapshot tests, should resolve with next marginaleffects version
   skip_if_not_installed("Formula")
   skip_if_not_installed("ggeffects")
   data(coffee_data, package = "ggeffects")

tests/testthat/test-estimate_means_mixed.R

@@ -1,3 +1,4 @@
+skip_on_cran()
 skip_if_not_installed("emmeans")
 skip_if_not_installed("marginaleffects")
 

tests/testthat/test-glmmTMB.R

@@ -1,3 +1,4 @@
+skip_on_cran()
 skip_if_not_installed("glmmTMB")
 skip_if_not_installed("emmeans")
 skip_if_not_installed("marginaleffects")