check_itemscale() accepts data frames

DESCRIPTION

@@ -1,7 +1,7 @@
 Type: Package
 Package: performance
 Title: Assessment of Regression Models Performance
-Version: 0.10.8.7
+Version: 0.10.8.8
 Authors@R: 
     c(person(given = "Daniel",
              family = "Lüdecke",

NAMESPACE

@@ -308,6 +308,7 @@ S3method(print,r2_nakagawa_by_group)
 S3method(print,r2_pseudo)
 S3method(print,test_likelihoodratio)
 S3method(print,test_performance)
+S3method(print_html,check_itemscale)
 S3method(print_html,compare_performance)
 S3method(print_html,test_performance)
 S3method(print_md,check_itemscale)

NEWS.md

@@ -5,6 +5,13 @@
 * `r2()` for models of class `glmmTMB` without random effects now returns the
   correct r-squared value for non-mixed models.
 
+* `check_itemscale()` now also accepts data frames as input. In this case,
+  `factor_index` must be specified, which must be a numeric vector of same
+  length as number of columns in `x`, where each element is the index of the
+  factor to which the respective column in `x`.
+
+* `check_itemscale()` gets a `print_html()` method.
+
 ## Bug fixes
 
 * Fixed issue in `binned_residuals()` for models with binary outcome, where

R/check_itemscale.R

@@ -2,11 +2,14 @@
 #' @name check_itemscale
 #'
 #' @description Compute various measures of internal consistencies
-#'   applied to (sub)scales, which items were extracted using
-#'   `parameters::principal_components()`.
+#' applied to (sub)scales, which items were extracted using
+#' `parameters::principal_components()`.
 #'
 #' @param x An object of class `parameters_pca`, as returned by
-#'   [`parameters::principal_components()`].
+#' [`parameters::principal_components()`], or a data frame.
+#' @param factor_index If `x` is a data frame, `factor_index` must be specified.
+#' It must be a numeric vector of same length as number of columns in `x`, where
+#' each element is the index of the factor to which the respective column in `x`.
 #'
 #' @return A list of data frames, with related measures of internal
 #'   consistencies of each subscale.
@@ -48,21 +51,53 @@
 #' X <- matrix(rnorm(1600), 100, 16)
 #' Z <- X %*% C
 #'
-#' pca <- principal_components(as.data.frame(Z), rotation = "varimax", n = 3)
+#' pca <- parameters::principal_components(
+#'   as.data.frame(Z),
+#'   rotation = "varimax",
+#'   n = 3
+#' )
 #' pca
 #' check_itemscale(pca)
+#'
+#' # as data frame
+#' check_itemscale(
+#'   as.data.frame(Z),
+#'   factor_index = parameters::closest_component(pca)
+#' )
 #' @export
-check_itemscale <- function(x) {
-  if (!inherits(x, "parameters_pca")) {
+check_itemscale <- function(x, factor_index = NULL) {
+  # check for valid input
+  if (!inherits(x, c("parameters_pca", "data.frame"))) {
     insight::format_error(
-      "`x` must be an object of class `parameters_pca`, as returned by `parameters::principal_components()`."
+      "`x` must be an object of class `parameters_pca`, as returned by `parameters::principal_components()`, or a data frame." # nolint
     )
   }
 
-  insight::check_if_installed("parameters")
+  # if data frame, we need `factor_index`
+  if (inherits(x, "data.frame") && !inherits(x, "parameters_pca")) {
+    if (is.null(factor_index)) {
+      insight::format_error("If `x` is a data frame, `factor_index` must be specified.")
+    }
+    if (!is.numeric(factor_index)) {
+      insight::format_error("`factor_index` must be numeric.")
+    }
+    if (length(factor_index) != ncol(x)) {
+      insight::format_error(
+        "`factor_index` must be of same length as number of columns in `x`.",
+        "Each element of `factor_index` must be the index of the factor to which the respective column in `x` belongs to." # nolint
+      )
+    }
+  }
 
-  dataset <- attributes(x)$dataset
-  subscales <- parameters::closest_component(x)
+  # assign data and factor index
+  if (inherits(x, "parameters_pca")) {
+    insight::check_if_installed("parameters")
+    dataset <- attributes(x)$dataset
+    subscales <- parameters::closest_component(x)
+  } else {
+    dataset <- x
+    subscales <- factor_index
+  }
 
   out <- lapply(sort(unique(subscales)), function(.subscale) {
     columns <- names(subscales)[subscales == .subscale]
@@ -123,3 +158,26 @@ print.check_itemscale <- function(x, digits = 2, ...) {
     zap_small = TRUE
   ))
 }
+
+
+#' @export
+print_html.check_itemscale <- function(x, digits = 2, ...) {
+  x <- lapply(seq_along(x), function(i) {
+    out <- x[[i]]
+    attr(out, "table_caption") <- sprintf(
+      "Component %i: Mean inter-item-correlation = %.3f, Cronbach's alpha = %.3f",
+      i,
+      attributes(out)$item_intercorrelation,
+      attributes(out)$cronbachs_alpha
+    )
+    out
+  })
+  insight::export_table(
+    x,
+    caption = "Description of (Sub-)Scales",
+    digits = digits,
+    format = "html",
+    missing = "<NA>",
+    zap_small = TRUE
+  )
+}

R/performance_score.R

@@ -52,7 +52,7 @@
 #' performance_score(model)
 #' }
 #' @export
 performance_score <- function(model, verbose = TRUE, ...) {
   # check special case
   if (inherits(model, c("logitor", "logitmfx", "probitmfx", "negbinirr", "negbinmfx", "poissonirr", "poissonmfx"))) {
     model <- model$fit
@@ -209,27 +209,25 @@
   pred_zi <- NULL
 
   tryCatch(
-    {
-      if (inherits(model, "MixMod")) {
-        pred <- stats::predict(model, type = "subject_specific")
-        pred_zi <- if (!is.null(model$gammas)) attr(pred, "zi_probs")
-      } else if (inherits(model, "glmmTMB")) {
-        pred <- stats::predict(model, type = "response")
-        pred_zi <- stats::predict(model, type = "zprob")
-      } else if (inherits(model, c("hurdle", "zeroinfl"))) {
-        pred <- stats::predict(model, type = "response")
-        pred_zi <- stats::predict(model, type = "zero")
-      } else if (inherits(model, c("clm", "clm2", "clmm"))) {
-        pred <- stats::predict(model)
-      } else if (all(inherits(model, c("stanreg", "lmerMod"), which = TRUE)) > 0) {
-        insight::check_if_installed("rstanarm")
-        pred <- colMeans(rstanarm::posterior_predict(model))
-      } else {
-        pred <- stats::predict(model, type = "response")
-      }
+    if (inherits(model, "MixMod")) {
+      pred <- stats::predict(model, type = "subject_specific")
+      pred_zi <- if (!is.null(model$gammas)) attr(pred, "zi_probs")
+    } else if (inherits(model, "glmmTMB")) {
+      pred <- stats::predict(model, type = "response")
+      pred_zi <- stats::predict(model, type = "zprob")
+    } else if (inherits(model, c("hurdle", "zeroinfl"))) {
+      pred <- stats::predict(model, type = "response")
+      pred_zi <- stats::predict(model, type = "zero")
+    } else if (inherits(model, c("clm", "clm2", "clmm"))) {
+      pred <- stats::predict(model)
+    } else if (all(inherits(model, c("stanreg", "lmerMod"), which = TRUE)) > 0) {
+      insight::check_if_installed("rstanarm")
+      pred <- colMeans(rstanarm::posterior_predict(model))
+    } else {
+      pred <- stats::predict(model, type = "response")
     },
     error = function(e) {
-      return(NULL)
+      NULL
     }
   )
 

tests/testthat/_snaps/windows/check_itemscale.md

@@ -0,0 +1,24 @@
+# check_itemscale
+
+    Code
+      print(out)
+    Output
+      # Description of (Sub-)ScalesComponent 1
+
+      Item | Missings | Mean |   SD | Skewness | Difficulty | Discrimination | alpha if deleted
+      -----------------------------------------------------------------------------------------
+      b    |        0 | 5.02 | 0.79 |    -0.04 |       0.84 |           0.06 |            -0.55
+      e    |        0 | 2.12 | 0.81 |    -0.22 |       0.35 |          -0.09 |            -0.03
+      f    |        0 | 2.00 | 0.82 |     0.00 |       0.33 |          -0.16 |             0.17
+
+      Mean inter-item-correlation = -0.046  Cronbach's alpha = -0.159
+      Component 2
+
+      Item | Missings | Mean |   SD | Skewness | Difficulty | Discrimination | alpha if deleted
+      -----------------------------------------------------------------------------------------
+      a    |        0 | 5.02 | 0.83 |    -0.04 |       0.84 |           0.21 |            -0.18
+      c    |        0 | 4.74 | 0.81 |     0.51 |       0.79 |          -0.04 |             0.41
+      d    |        0 | 2.07 | 0.79 |    -0.13 |       0.34 |           0.13 |             0.04
+
+      Mean inter-item-correlation = 0.067  Cronbach's alpha = 0.178
+

tests/testthat/test-check_itemscale.R

@@ -1,4 +1,4 @@
-test_that("check_convergence", {
+test_that("check_itemscale", {
   skip_if_not_installed("parameters")
 
   set.seed(123)
@@ -25,4 +25,31 @@ test_that("check_convergence", {
     tolerance = 1e-4,
     ignore_attr = TRUE
   )
+  expect_snapshot(print(out), variant = "windows")
+  comp <- parameters::closest_component(pca)
+  out2 <- check_itemscale(d, comp)
+  expect_equal(
+    out[[1]]$Mean,
+    out2[[1]]$Mean,
+    tolerance = 1e-4,
+    ignore_attr = TRUE
+  )
+  expect_equal(
+    out[[1]]$Difficulty,
+    out2[[1]]$Difficulty,
+    tolerance = 1e-4,
+    ignore_attr = TRUE
+  )
+  expect_error(
+    check_itemscale(d),
+    regex = "If `x` is a data"
+  )
+  expect_error(
+    check_itemscale(d, factor_index = 1:8),
+    regex = "`factor_index` must be of same"
+  )
+  expect_error(
+    check_itemscale(d, factor_index = factor(comp)),
+    regex = "`factor_index` must be numeric."
+  )
 })

tests/testthat/test-glmmPQL.R

@@ -11,5 +11,5 @@ test_that("r2", {
     random = ~ 1 | species,
     family = "quasibinomial", data = example_dat
   )
-  expect_message(performance_score(mn), regex = "Cant calculate")
+  expect_message(performance_score(mn), regex = "Can't calculate")
 })