diff --git a/r-package/grf/R/multi_arm_causal_forest.R b/r-package/grf/R/multi_arm_causal_forest.R
index c8ef4fdb9..fc5b86d59 100644
--- a/r-package/grf/R/multi_arm_causal_forest.R
+++ b/r-package/grf/R/multi_arm_causal_forest.R
@@ -132,7 +132,7 @@
 #' abline(0, -1.5, col = "red")
 #' legend("topleft", c("B - A", "C - A"), col = c("black", "blue"), pch = 19)
 #'
-#' # The average treatment effect of the arms with "A" as baseline.
+#' # A doubly robust estimate (AIPW) of the average treatment effect of the arms.
 #' average_treatment_effect(mc.forest)
 #'
 #' # The conditional response surfaces mu_k(X) for a single outcome can be reconstructed from
diff --git a/r-package/grf/man/multi_arm_causal_forest.Rd b/r-package/grf/man/multi_arm_causal_forest.Rd
index 68815a5ef..7fd10f566 100644
--- a/r-package/grf/man/multi_arm_causal_forest.Rd
+++ b/r-package/grf/man/multi_arm_causal_forest.Rd
@@ -180,7 +180,7 @@ points(X[, 2], tau.hat[, "C - A"], col = "blue")
 abline(0, -1.5, col = "red")
 legend("topleft", c("B - A", "C - A"), col = c("black", "blue"), pch = 19)
 
-# The average treatment effect of the arms with "A" as baseline.
+# A doubly robust estimate (AIPW) of the average treatment effect of the arms.
 average_treatment_effect(mc.forest)
 
 # The conditional response surfaces mu_k(X) for a single outcome can be reconstructed from