diff --git a/NEWS.md b/NEWS.md
index 8c4ad86d..ec5b4471 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -1,5 +1,8 @@
 # hardhat (development version)
 
+* Added more documentation about importance and frequency weights in
+  `?importance_weights()` and `?frequency_weights()` (#214).
+
 # hardhat 1.2.0
 
 * We have reverted the change made in hardhat 1.0.0 that caused recipe
diff --git a/R/case-weights.R b/R/case-weights.R
index 3b8f1652..762c73ac 100644
--- a/R/case-weights.R
+++ b/R/case-weights.R
@@ -8,6 +8,15 @@
 #' are supplied as a non-negative double vector, where fractional values are
 #' allowed.
 #'
+#' @details
+#' Importance weights focus on how much each row of the data set should
+#' influence model estimation. These can be based on data or arbitrarily set to
+#' achieve some goal.
+#'
+#' In tidymodels, importance weights only affect the model estimation and
+#' _supervised_ recipes steps. They are not used with yardstick functions for
+#' calculating measures of model performance.
+#'
 #' @param x A double vector.
 #'
 #' @return A new importance weights vector.
@@ -117,6 +126,14 @@ vec_ptype_abbr.hardhat_importance_weights <- function(x, ...) {
 #' are supplied as a non-negative integer vector, where only whole numbers are
 #' allowed.
 #'
+#' @details
+#' Frequency weights are integers that denote how many times a particular row of
+#' the data has been observed. They help compress redundant rows into a single
+#' entry.
+#'
+#' In tidymodels, frequency weights are used for all parts of the preprocessing,
+#' model fitting, and performance estimation operations.
+#'
 #' @param x An integer vector.
 #'
 #' @return A new frequency weights vector.
diff --git a/man/frequency_weights.Rd b/man/frequency_weights.Rd
index a830251b..6bbd1f1c 100644
--- a/man/frequency_weights.Rd
+++ b/man/frequency_weights.Rd
@@ -20,6 +20,14 @@ to compactly repeat an observation a set number of times. Frequency weights
 are supplied as a non-negative integer vector, where only whole numbers are
 allowed.
 }
+\details{
+Frequency weights are integers that denote how many times a particular row of
+the data has been observed. They help compress redundant rows into a single
+entry.
+
+In tidymodels, frequency weights are used for all parts of the preprocessing,
+model fitting, and performance estimation operations.
+}
 \examples{
 # Record that the first observation has 10 replicates, the second has 12
 # replicates, and so on
diff --git a/man/importance_weights.Rd b/man/importance_weights.Rd
index 56030b40..1a71613e 100644
--- a/man/importance_weights.Rd
+++ b/man/importance_weights.Rd
@@ -20,6 +20,15 @@ to apply a context dependent weight to your observations. Importance weights
 are supplied as a non-negative double vector, where fractional values are
 allowed.
 }
+\details{
+Importance weights focus on how much each row of the data set should
+influence model estimation. These can be based on data or arbitrarily set to
+achieve some goal.
+
+In tidymodels, importance weights only affect the model estimation and
+\emph{supervised} recipes steps. They are not used with yardstick functions for
+calculating measures of model performance.
+}
 \examples{
 importance_weights(c(1.5, 2.3, 10))
 }