Permalink
Browse files

Tweaks to stat_smooth docs, thanks to @wch

  • Loading branch information...
1 parent 62927fb commit cbd7401abfa964a83b728196c3ed4be72a736412 @hadley hadley committed Dec 20, 2011
Showing with 9 additions and 4 deletions.
  1. +4 −2 R/stat-smooth.r
  2. +5 −2 man/stat_smooth.Rd
View
@@ -5,8 +5,10 @@
#' Calculation is performed by the (currently undocumented)
#' \code{predictdf} generic function and its methods. For most methods
#' the confidence bounds are computed using the \code{\link{predict}}
-#' method - the exception is for \code{loess} which uses a t-based
-#' appromxiation
+#' method - the exceptions are \code{loess} which uses a t-based
+#' approximation, and for \code{glm} where the normal confidence interval
+#' is constructed on the link scale, and then back-transformed to the response
+#' scale.
#'
#' @param method smoothing method (function) to use, eg. lm, glm, gam, loess,
#' rlm
View
@@ -45,8 +45,11 @@
Calculation is performed by the (currently undocumented)
\code{predictdf} generic function and its methods. For
most methods the confidence bounds are computed using the
- \code{\link{predict}} method - the exception is for
- \code{loess} which uses a t-based appromxiation
+ \code{\link{predict}} method - the exceptions are
+ \code{loess} which uses a t-based approximation, and for
+ \code{glm} where the normal confidence interval is
+ constructed on the link scale, and then back-transformed
+ to the response scale.
}
\examples{
c <- ggplot(mtcars, aes(qsec, wt))

0 comments on commit cbd7401

Please sign in to comment.