diff --git a/NEWS.md b/NEWS.md
index 5475776cc..a4fd3df0f 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -23,6 +23,8 @@
 * `performance_score()` should no longer fail for models where scoring rules
   can't be calculated. Instead, an informative message is returned.
 
+* `check_outliers()` now properly accept the `percentage_central` argument when using the `"mcd"` method.
+
 # performance 0.10.8
 
 ## Changes
diff --git a/R/check_outliers.R b/R/check_outliers.R
index 2a1d185d3..e01b20e12 100644
--- a/R/check_outliers.R
+++ b/R/check_outliers.R
@@ -160,6 +160,9 @@
 #' the data (by default, 66\%), before computing the Mahalanobis Distance. This
 #' is deemed to be a more robust method of identifying and removing outliers
 #' than regular Mahalanobis distance.
+#' This method has a `percentage_central` argument that allows specifying
+#' the breakdown point (0.75, the default, is recommended by Leys et al. 2018,
+#' but a commonly used alternative is 0.50).
 #'
 #'  - **Invariant Coordinate Selection (ICS)**:
 #'  The outlier are detected using ICS, which by default uses an alpha threshold
@@ -1098,8 +1101,8 @@ check_outliers.data.frame <- function(x,
     out <- c(out, .check_outliers_mcd(
       x,
       threshold = thresholds$mcd,
-      percentage_central = 0.66,
-      ID.names = ID.names
+      ID.names = ID.names,
+      ...
     ))
 
     count.table <- datawizard::data_filter(
@@ -1726,7 +1729,7 @@ check_outliers.metabin <- check_outliers.metagen
 
 .check_outliers_mcd <- function(x,
                                 threshold = stats::qchisq(p = 1 - 0.001, df = ncol(x)),
-                                percentage_central = 0.50,
+                                percentage_central = 0.75,
                                 ID.names = NULL) {
   out <- data.frame(Row = seq_len(nrow(x)))
 
diff --git a/man/check_outliers.Rd b/man/check_outliers.Rd
index 74c992b6f..9c2a046b4 100644
--- a/man/check_outliers.Rd
+++ b/man/check_outliers.Rd
@@ -189,6 +189,9 @@ calculates the mean and covariance matrix based on the most central subset of
 the data (by default, 66\\%), before computing the Mahalanobis Distance. This
 is deemed to be a more robust method of identifying and removing outliers
 than regular Mahalanobis distance.
+This method has a \code{percentage_central} argument that allows specifying
+the breakdown point (0.75, the default, is recommended by Leys et al. 2018,
+but a commonly used alternative is 0.50).
 \item \strong{Invariant Coordinate Selection (ICS)}:
 The outlier are detected using ICS, which by default uses an alpha threshold
 of 0.025 (corresponding to the 2.5\\% most extreme observations) as a cut-off