diff --git a/R/parser.R b/R/parser.R
index e7f0a8f..ce2ef3a 100644
--- a/R/parser.R
+++ b/R/parser.R
@@ -28,6 +28,7 @@ collector_find <- function(name) {
 #' @param .return_problems Whether to hide the `problems` tibble from the output
 #' @keywords internal
 #' @export
+#' @return a parsed vector
 #' @examples
 #' x <- c("1", "2", "3", "NA")
 #' parse_vector(x, col_integer())
@@ -65,6 +66,7 @@ parse_vector <- function(x, collector, na = c("", "NA"), locale = default_locale
 #'     each field before parsing it?
 #' @inheritParams parse_vector
 #' @family parsers
+#' @return a parsed vector
 #' @examples
 #' parse_integer(c("1", "2", "3"))
 #' parse_double(c("1", "2", "3.123"))
@@ -129,11 +131,7 @@ col_character <- function() {
     collector("character")
 }
 
-#' Skip a column
-#'
-#' Use this function to ignore a column when parsing.
-#' To skip all columns not otherwise specified, use [cols_only()].
-#'
+#' @rdname cols
 #' @family parsers
 #' @export
 col_skip <- function() {
@@ -150,6 +148,7 @@ col_skip <- function() {
 #' @return A numeric vector (double) of parsed numbers.
 #' @family parsers
 #' @export
+#' @return a parsed vector
 #' @examples
 #' ## These all return 1000
 #' parse_number("$1,000") ## leading `$` and grouping character `,` ignored
@@ -187,6 +186,7 @@ col_number <- function() {
 #' @param guess_integer If `TRUE`, guess integer types for whole numbers, if
 #'   `FALSE` guess numeric type for all numbers.
 #' @family parsers
+#' @return a parsed vector
 #' @export
 #' @examples
 #' # Logical vectors
@@ -232,6 +232,7 @@ guess_parser <- function(x, locale = default_locale(), guess_integer = FALSE, na
 #'   is included in the levels of the constructed factor.
 #'
 #' @inheritParams parse_atomic
+#' @return a parsed vector
 #' @family parsers
 #' @export
 #' @examples
@@ -464,6 +465,7 @@ col_time <- function(format = "") {
 #'   ASCII? This is useful if you're dealing with ASCII data where the correct
 #'   spellings have been lost. Requires the \pkg{stringi} package.
 #' @export
+#' @return a list / S3 object representing the locale information
 #' @examples
 #' locale()
 #' locale("fr")
@@ -569,6 +571,7 @@ check_encoding <- function(x) {
 #' @param mon,mon_ab Full and abbreviated month names.
 #' @param day,day_ab Full and abbreviated week day names. Starts with Sunday.
 #' @param am_pm Names used for AM and PM.
+#' @return a list / S3 object representing data time
 #' @export
 #' @examples
 #' date_names_lang("en")
@@ -649,6 +652,7 @@ is.date_names <- function(x) inherits(x, "date_names")
 #' @param .default Any named columns not explicitly overridden in `...`
 #'   will be read with this column type.
 #' @export
+#' @return a list / S3 object representing column specification
 #' @examples
 #' cols(a = col_integer())
 #' cols_only(a = col_integer())
@@ -722,6 +726,7 @@ is.col_spec <- function(x) inherits(x, "col_spec")
 #' @keywords internal
 #' @examples
 #' as.col_spec("cccnnn")
+#' @return a list / S3 object representing column specification
 #' @export
 as.col_spec <- function(x) UseMethod("as.col_spec")
 #' @export
diff --git a/R/type_convert.R b/R/type_convert.R
index ff0ef2c..3537a21 100644
--- a/R/type_convert.R
+++ b/R/type_convert.R
@@ -14,6 +14,7 @@
 #' @inheritParams parse_guess
 #' @note `type_convert()` removes a 'spec' attribute (if it presents).
 #' @export
+#' @return A data frame
 #' @examples
 #' df <- data.frame(
 #'   x = as.character(runif(10)),
diff --git a/man/as.col_spec.Rd b/man/as.col_spec.Rd
index 3d3f4b4..597263c 100644
--- a/man/as.col_spec.Rd
+++ b/man/as.col_spec.Rd
@@ -9,6 +9,9 @@ as.col_spec(x)
 \arguments{
 \item{x}{Input object}
 }
+\value{
+a list / S3 object representing column specification
+}
 \description{
 This is most useful for generating a specification using the short form
 }
diff --git a/man/col_skip.Rd b/man/col_skip.Rd
deleted file mode 100644
index e32e849..0000000
--- a/man/col_skip.Rd
+++ /dev/null
@@ -1,23 +0,0 @@
-% Generated by roxygen2: do not edit by hand
-% Please edit documentation in R/parser.R
-\name{col_skip}
-\alias{col_skip}
-\title{Skip a column}
-\usage{
-col_skip()
-}
-\description{
-Use this function to ignore a column when parsing.
-To skip all columns not otherwise specified, use \code{\link[=cols_only]{cols_only()}}.
-}
-\seealso{
-Other parsers: 
-\code{\link{cols}()},
-\code{\link{parse_datetime}()},
-\code{\link{parse_factor}()},
-\code{\link{parse_guess}()},
-\code{\link{parse_logical}()},
-\code{\link{parse_number}()},
-\code{\link{parse_vector}()}
-}
-\concept{parsers}
diff --git a/man/cols.Rd b/man/cols.Rd
index 2b505a1..97ad0d6 100644
--- a/man/cols.Rd
+++ b/man/cols.Rd
@@ -1,10 +1,13 @@
 % Generated by roxygen2: do not edit by hand
 % Please edit documentation in R/parser.R
-\name{cols}
+\name{col_skip}
+\alias{col_skip}
 \alias{cols}
 \alias{cols_only}
 \title{Create column specification}
 \usage{
+col_skip()
+
 cols(..., .default = col_guess())
 
 cols_only(...)
@@ -19,6 +22,9 @@ the column names exactly.}
 \item{.default}{Any named columns not explicitly overridden in \code{...}
 will be read with this column type.}
 }
+\value{
+a list / S3 object representing column specification
+}
 \description{
 \code{cols()} includes all columns in the input data, guessing the column types
 as the default. \code{cols_only()} includes only the columns you explicitly
@@ -67,7 +73,14 @@ t3
 }
 \seealso{
 Other parsers: 
-\code{\link{col_skip}()},
+\code{\link{parse_datetime}()},
+\code{\link{parse_factor}()},
+\code{\link{parse_guess}()},
+\code{\link{parse_logical}()},
+\code{\link{parse_number}()},
+\code{\link{parse_vector}()}
+
+Other parsers: 
 \code{\link{parse_datetime}()},
 \code{\link{parse_factor}()},
 \code{\link{parse_guess}()},
diff --git a/man/date_names.Rd b/man/date_names.Rd
index 3a1f3ea..d532bb7 100644
--- a/man/date_names.Rd
+++ b/man/date_names.Rd
@@ -23,6 +23,9 @@ date_names_langs()
 e.g. \code{"en"} for American English. See \code{date_names_langs()}
 for a complete list of available locales.}
 }
+\value{
+a list / S3 object representing data time
+}
 \description{
 When parsing dates, you often need to know how weekdays of the week and
 months are represented as text. This pair of functions allows you to either
diff --git a/man/locale.Rd b/man/locale.Rd
index 5192613..497d180 100644
--- a/man/locale.Rd
+++ b/man/locale.Rd
@@ -50,6 +50,9 @@ DST. It is \emph{not} Eastern Standard Time. It's better to use
 ASCII? This is useful if you're dealing with ASCII data where the correct
 spellings have been lost. Requires the \pkg{stringi} package.}
 }
+\value{
+a list / S3 object representing the locale information
+}
 \description{
 A locale object tries to capture all the defaults that can vary between
 countries. You set the locale in once, and the details are automatically
diff --git a/man/parse_atomic.Rd b/man/parse_atomic.Rd
index 937b124..916bcac 100644
--- a/man/parse_atomic.Rd
+++ b/man/parse_atomic.Rd
@@ -68,6 +68,9 @@ each field before parsing it?}
 
 \item{.return_problems}{Whether to hide the \code{problems} tibble from the output}
 }
+\value{
+a parsed vector
+}
 \description{
 Use \verb{parse_*()} if you have a character vector you want to parse.
 }
@@ -89,7 +92,6 @@ parse_double(x, na = "-")
 \seealso{
 Other parsers: 
 \code{\link{col_skip}()},
-\code{\link{cols}()},
 \code{\link{parse_datetime}()},
 \code{\link{parse_factor}()},
 \code{\link{parse_guess}()},
diff --git a/man/parse_datetime.Rd b/man/parse_datetime.Rd
index fd243bf..add1c31 100644
--- a/man/parse_datetime.Rd
+++ b/man/parse_datetime.Rd
@@ -195,7 +195,6 @@ parse_datetime("1979-10-14T1010", locale = locale(tz = ""))
 \seealso{
 Other parsers: 
 \code{\link{col_skip}()},
-\code{\link{cols}()},
 \code{\link{parse_factor}()},
 \code{\link{parse_guess}()},
 \code{\link{parse_logical}()},
diff --git a/man/parse_factor.Rd b/man/parse_factor.Rd
index 41f2f35..b58ecb8 100644
--- a/man/parse_factor.Rd
+++ b/man/parse_factor.Rd
@@ -44,6 +44,9 @@ each field before parsing it?}
 
 \item{.return_problems}{Whether to hide the \code{problems} tibble from the output}
 }
+\value{
+a parsed vector
+}
 \description{
 \code{parse_factor()} is similar to \code{\link[=factor]{factor()}}.
 }
@@ -71,7 +74,6 @@ parse_factor(x, levels = animals)
 \seealso{
 Other parsers: 
 \code{\link{col_skip}()},
-\code{\link{cols}()},
 \code{\link{parse_datetime}()},
 \code{\link{parse_guess}()},
 \code{\link{parse_logical}()},
diff --git a/man/parse_guess.Rd b/man/parse_guess.Rd
index 396e31e..0d5e225 100644
--- a/man/parse_guess.Rd
+++ b/man/parse_guess.Rd
@@ -36,6 +36,9 @@ each field before parsing it?}
 
 \item{.return_problems}{Whether to hide the \code{problems} tibble from the output}
 }
+\value{
+a parsed vector
+}
 \description{
 \code{parse_guess()} returns the parser vector. This function uses a number of heuristics
 to determine which type of vector is "best". Generally they try to err of
@@ -59,7 +62,6 @@ parse_guess(c("2010-10-10"))
 \seealso{
 Other parsers: 
 \code{\link{col_skip}()},
-\code{\link{cols}()},
 \code{\link{parse_datetime}()},
 \code{\link{parse_factor}()},
 \code{\link{parse_logical}()},
diff --git a/man/parse_number.Rd b/man/parse_number.Rd
index c5fbd3f..38f1069 100644
--- a/man/parse_number.Rd
+++ b/man/parse_number.Rd
@@ -34,6 +34,8 @@ each field before parsing it?}
 }
 \value{
 A numeric vector (double) of parsed numbers.
+
+a parsed vector
 }
 \description{
 This parses the first number it finds, dropping any non-numeric characters
@@ -59,7 +61,6 @@ parse_number(c("1", "2", "3", "NA", "Nothing"), na = c("NA", "Nothing"))
 \seealso{
 Other parsers: 
 \code{\link{col_skip}()},
-\code{\link{cols}()},
 \code{\link{parse_datetime}()},
 \code{\link{parse_factor}()},
 \code{\link{parse_guess}()},
diff --git a/man/parse_vector.Rd b/man/parse_vector.Rd
index 623e1ed..2fae54c 100644
--- a/man/parse_vector.Rd
+++ b/man/parse_vector.Rd
@@ -20,6 +20,9 @@ parse_vector(
 
 \item{.return_problems}{Whether to hide the \code{problems} tibble from the output}
 }
+\value{
+a parsed vector
+}
 \description{
 Parse a character vector.
 }
@@ -31,7 +34,6 @@ parse_vector(x, col_double())
 \seealso{
 Other parsers: 
 \code{\link{col_skip}()},
-\code{\link{cols}()},
 \code{\link{parse_datetime}()},
 \code{\link{parse_factor}()},
 \code{\link{parse_guess}()},
diff --git a/man/type_convert.Rd b/man/type_convert.Rd
index 4d570ce..eeac72b 100644
--- a/man/type_convert.Rd
+++ b/man/type_convert.Rd
@@ -39,6 +39,9 @@ names.}
 
 \item{verbose}{whether to print messages}
 }
+\value{
+A data frame
+}
 \description{
 This is useful if you need to do some manual munging - you can read the
 columns in as character, clean it up with (e.g.) regular expressions and