Nothing
####################################################
# In order to let roxygen2() create all man pages, #
# you must run the following command: #
# roxygen2::roxygenise() #
####################################################
#' Datatable of the Messwiederholung Example for ANOVA
#'
#' @details Variables in the dataset:
#' \itemize{
#' \item Name. The first name of the probands.
#' \item t1. Measure at time 1
#' \item t2. Measure at time 2
#' \item t3. Measure at time 3
#' }
#'
#' @title Datatable of the Messwiederholung Example for ANOVA
#' @docType data
#' @keywords datasets
#' @name Messwiederholung
#' @usage data(Messwiederholung)
#' @format A data frame with 200 observations in 4 variables
#' @source \url{https://www.produnis.de/R/}
NULL
#' Datatable of the SuperMario Example for Friedman-ANOVA
#'
#' @details Variables in the dataset:
#' \itemize{
#' \item Name. The characters' name
#' \item Alter. The characters' age in years
#' \item Kingdom. The characters' home
#' \item Geschlecht. The characters' gender (männlich = male, weiblich = female)
#' \item BadGuy. Whether the character is a bad guy, logical
#' \item t1. Measure at time 1
#' \item t2. Measure at time 2
#' \item t3. Measure at time 3
#' }
#'
#' @title Datatable of the SuperMario Example for Friedman-ANOVA
#' @docType data
#' @keywords datasets
#' @name MarioANOVA
#' @usage data(MarioANOVA)
#' @format A data frame with 47 observations in 8 variables
#' @source \url{https://www.produnis.de/R/}
NULL
#' Datatable of the Faktorenbogen Example for factor analysis
#'
#' @details Variables in the dataset:
#' \itemize{
#' \item gender. a factor with levels \code{female} \code{male} \code{other}, giving the proband's gender
#' \item age. a numeric vector of proband's age in years
#' \item A. Item A of the questionnaire, numeric
#' \item B. Item B of the questionnaire, numeric
#' \item C. Item C of the questionnaire, numeric
#' \item D. Item D of the questionnaire, numeric
#' \item E. Item E of the questionnaire, numeric
#' \item F. Item F of the questionnaire, numeric
#' \item G. Item G of the questionnaire, numeric
#' \item H. Item H of the questionnaire, numeric
#' \item I. Item I of the questionnaire, numeric
#' \item J. Item J of the questionnaire, numeric
#' \item K. Item K of the questionnaire, numeric
#' \item L. Item L of the questionnaire, numeric
#' }
#'
#' @title Datatable of the Faktorenbogen Example for factor analysis
#' @docType data
#' @keywords datasets
#' @name Faktorenbogen
#' @usage data(Faktorenbogen)
#' @format A data frame with 150 observations in 14 variables
#' @source \url{https://www.produnis.de/R/}
NULL
#' Datatable of the epa Example
#'
#' @details Variables in the dataset:
#' \itemize{
#' \item sex. a factor with levels \code{m} \code{w} \code{d}, giving the proband's sex
#' \item age. a numeric vector
#' \item cms. a numeric vector
#' \item risk. a dichotome vector, 0 = not at risk, 1 = at risk
#' \item expert. a dichotome vector of expert's decision, 0 = not at risk, 1 = at risk
#' \item decu. a dichotome vector, 0 = no decubitus, 1 = decubitus
#' }
#'
#' @title Datatable of the epa Example
#' @docType data
#' @keywords datasets
#' @name epa
#' @usage data(epa)
#' @format A data frame with 620 observations in 6 variables
#' @source \url{https://www.produnis.de/R/}
NULL
#' Dataset of a work sampling study
#'
#' @details Variables in the dataset:
#' \itemize{
#' \item day. a vector, giving the number of the observation day
#' \item time. a factor giving the time of observation
#' \item ward. a factor giving the ward under observation
#' \item qual. a factor giving the qualification of the nurse
#' \item category. a factor of qualification categories
#' \item action. a factor giving the observed action
#' }
#' @title Dataset of a work sampling study
#' @docType data
#' @keywords datasets
#' @name mma
#' @usage data(mma)
#' @format A data frame with 9768 observations in 6 variables.
#' @source \url{https://www.produnis.de/R/}
NULL
#' Dataset of the German Nachtwachen study
#'
#' @title Dataset of the German Nachtwachen study
#' @docType data
#' @keywords datasets
#' @name Nachtwachen
#' @usage data(Nachtwachen)
#' @format A data frame with 276 observations in 37 variables.
#' @source \url{https://www.produnis.de/R/}
NULL
#' Dataset of the German Nachtwachen study, labelled version
#'
#' @title Dataset of the German Nachtwachen study with labelled variables
#' @docType data
#' @keywords datasets
#' @name nw
#' @aliases nw_labelled
#' @usage data(nw)
#' @format A data frame with 276 observations in 37 variables.
#' @source \url{https://www.produnis.de/R/}
NULL
#' Datatable of an Ordinal Sample
#'
#' @title Datatable of an Ordinal Sample
#' @docType data
#' @keywords datasets
#' @name OrdinalSample
#' @aliases ordinalSample
#' @usage data(OrdinalSample)
#' @format A data frame with 415 observations in 4 variables.
#' @source \url{https://www.produnis.de/R/}
#' @details Variables in the dataset:
#' \itemize{
#' \item Konflikt. a numeric vector giving the potential of conflicts.
#' \item Zufriedenh. a numeric vector giving the satisfaction of workers
#' \item Geschlecht. a factor of proband's sex, 1 = male, 2=female
#' \item Stimmung. an ordinal factor of proband's mood
#' }
NULL
#' This is the dataset of the PF8 example.
#'
#' @title Dataset of the PF8 example.
#' @docType data
#' @keywords datasets
#' @name pf8
#' @usage data(pf8)
#' @format A data frame with 731 observations in 16 variables.
#' @source \url{https://www.produnis.de/R/}
NULL
#' Matrix of Pflegeberufe by Isfort et al. 2018
#'
#' @title Matrix of Pflegeberufe by Isfort et al. 2018
#' @docType data
#' @keywords datasets
#' @name Pflegeberufe
#' @usage data(Pflegeberufe)
#' @format A matrix with 9 cols (years) and 5 rows (nursing profession).
#' @author Isfort et al. 2018 (Pflegethermometer)
#' @source \url{https://www.produnis.de/R/}
NULL
Any scripts or data that you put into this service are public.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.