diff --git a/R/hop-index.R b/R/hop-index.R
index eccd7b9..b17aa4b 100644
--- a/R/hop-index.R
+++ b/R/hop-index.R
@@ -63,7 +63,7 @@
 #' hop_index(i, i, starts, stops, ~.x)
 #'
 #' # ---------------------------------------------------------------------------
-#' # Non-existant dates with `lubridate::months()`
+#' # Non-existent dates with `lubridate::months()`
 #'
 #' # Imagine you want to compute a 1 month rolling average on this
 #' # irregular daily data.
@@ -81,7 +81,7 @@
 #' )
 #'
 #' # This is because when you actually compute the `.i - .before` sequence,
-#' # you hit non-existant dates. i.e. `"2019-03-29" - months(1)` doesn't exist.
+#' # you hit non-existent dates. i.e. `"2019-03-29" - months(1)` doesn't exist.
 #' i - months(1)
 #'
 #' # To get around this, lubridate provides `add_with_rollback()`,
diff --git a/R/slide-index2.R b/R/slide-index2.R
index 2a94d45..d63ec54 100644
--- a/R/slide-index2.R
+++ b/R/slide-index2.R
@@ -1,4 +1,4 @@
-#' Slide along multiples inputs simultaneously relative to an index
+#' Slide along multiple inputs simultaneously relative to an index
 #'
 #' `slide_index2()` and `pslide_index()` represent the combination
 #' of [slide2()] and [pslide()] with [slide_index()], allowing you to iterate
diff --git a/R/summary-index.R b/R/summary-index.R
index c59368a..fe3ba76 100644
--- a/R/summary-index.R
+++ b/R/summary-index.R
@@ -63,6 +63,30 @@
 #'
 #' # Only evaluate the sum on windows that have the potential to be complete
 #' slide_index_sum(x, i, before = 2, after = 1, complete = TRUE)
+#'
+#' # ---------------------------------------------------------------------------
+#' # Panel data (grouped rolling operations)
+#'
+#' # In Stata, the equivalent of grouped rolling means is:
+#' #   rangestat (mean) roll_mean = sales, int(date -2 0) by(firm)
+#'
+#' # slider + dplyr replicates this pattern with group_by():
+#' df <- data.frame(
+#'   firm = rep(c("A", "B"), each = 4),
+#'   date = as.Date(c(
+#'     "2020-01-01", "2020-01-02", "2020-01-05", "2020-01-06",
+#'     "2020-01-01", "2020-01-03", "2020-01-04", "2020-01-07"
+#'   )),
+#'   sales = c(10, 12, 15, 13, 20, 22, 18, 25)
+#' )
+#'
+#' # 3-day rolling mean within each firm, respecting irregular date gaps
+#' df$roll_mean <- unlist(
+#'   lapply(split(df, df$firm), function(g) {
+#'     slide_index_mean(g$sales, g$date, before = 2)
+#'   })
+#' )
+#' df
 slide_index_sum <- function(
   x,
   i,
diff --git a/man/hop_index.Rd b/man/hop_index.Rd
index 39e503a..8890ff3 100644
--- a/man/hop_index.Rd
+++ b/man/hop_index.Rd
@@ -125,7 +125,7 @@ stops <- vec_c(inner - 1, last_stop)
 hop_index(i, i, starts, stops, ~.x)
 
 # ---------------------------------------------------------------------------
-# Non-existant dates with `lubridate::months()`
+# Non-existent dates with `lubridate::months()`
 
 # Imagine you want to compute a 1 month rolling average on this
 # irregular daily data.
@@ -143,7 +143,7 @@ with_options(
 )
 
 # This is because when you actually compute the `.i - .before` sequence,
-# you hit non-existant dates. i.e. `"2019-03-29" - months(1)` doesn't exist.
+# you hit non-existent dates. i.e. `"2019-03-29" - months(1)` doesn't exist.
 i - months(1)
 
 # To get around this, lubridate provides `add_with_rollback()`,
diff --git a/man/slide_index2.Rd b/man/slide_index2.Rd
index 7eb5c45..82088d3 100644
--- a/man/slide_index2.Rd
+++ b/man/slide_index2.Rd
@@ -17,7 +17,7 @@
 \alias{pslide_index_chr}
 \alias{pslide_index_dfr}
 \alias{pslide_index_dfc}
-\title{Slide along multiples inputs simultaneously relative to an index}
+\title{Slide along multiple inputs simultaneously relative to an index}
 \usage{
 slide_index2(.x, .y, .i, .f, ..., .before = 0L, .after = 0L, .complete = FALSE)
 
diff --git a/man/summary-index.Rd b/man/summary-index.Rd
index e1ed27a..7d4af46 100644
--- a/man/summary-index.Rd
+++ b/man/summary-index.Rd
@@ -183,6 +183,30 @@ slide_index_mean(x, i, before = 2)
 
 # Only evaluate the sum on windows that have the potential to be complete
 slide_index_sum(x, i, before = 2, after = 1, complete = TRUE)
+
+# ---------------------------------------------------------------------------
+# Panel data (grouped rolling operations)
+
+# In Stata, the equivalent of grouped rolling means is:
+#   rangestat (mean) roll_mean = sales, int(date -2 0) by(firm)
+
+# slider + dplyr replicates this pattern with group_by():
+df <- data.frame(
+  firm = rep(c("A", "B"), each = 4),
+  date = as.Date(c(
+    "2020-01-01", "2020-01-02", "2020-01-05", "2020-01-06",
+    "2020-01-01", "2020-01-03", "2020-01-04", "2020-01-07"
+  )),
+  sales = c(10, 12, 15, 13, 20, 22, 18, 25)
+)
+
+# 3-day rolling mean within each firm, respecting irregular date gaps
+df$roll_mean <- unlist(
+  lapply(split(df, df$firm), function(g) {
+    slide_index_mean(g$sales, g$date, before = 2)
+  })
+)
+df
 }
 \seealso{
 \code{\link[=slide_sum]{slide_sum()}}