summarize_numeric: Produce Summary Tables for Data Sets

Description

The function produces summary tables for continuous variables. The obtained tables can be used directly in R, with LaTeX and HTML (by using the xtable function) or Markdown (e.g. by using the function kable).

Usage

summarize_numeric(data,
    variables = names(data), variable.labels = labels, labels = NULL,
    group = NULL, test = !is.null(group), colnames = NULL,
    digits = 2, digits.pval = 3, smallest.pval = 0.001,
    sep = !is.null(group), sanitize = TRUE,
    drop = TRUE, show.NAs = any(is.na(data[, variables])),
    ## additional specific arguments
    count = TRUE, mean_sd = TRUE, quantiles = TRUE,
    incl_outliers = TRUE, ...)

Value

A special data.frame with additional class summary

containing the computed statistics is returned from function

summarize. Addtional attributes required for the

xtable.summary or print.xtable.summary

function are contained as attributes. These are extracted using the function get_option.

Arguments

data: data set to be used.
variables: variables that should be included in the table. For details see summarize.
variable.labels, labels: labels for the variables. For details see summarize.
group: character specifying a grouping factor. For details see summarize.
test: logical or charachter specifying test for group differences. For details see summarize.
colnames: a vector of character strings of appropriate length. For details see summarize.
digits: number of digits to round to. Per default all values are rounded to two digits.
digits.pval: number of significant digits used for p-values.
smallest.pval: determines the smallest p-value to be printed exactly. For details see summarize.
sep: logical (default: TRUE if grouping specified, FALSE otherwise). Determines whether separators (lines) should be added after each variable.
sanitize: logical (default: TRUE) or a sanitizing function. For details see summarize.
drop: logical (default: TRUE). Determines whether variables, which contain only missing values are dropped from the table.
show.NAs: logical. Determines if the number of missings (NAs) is displayed as a separate column. Per default, show.NAs is TRUE if there are any missings in the variables to be displayed (and FALSE if not).
count: (logical) indicator if number of complete cases ("n") should be included in the table (default: TRUE).
mean_sd: (logical) indicator if mean and standard deviation should be included in the table (default: TRUE).
quantiles: (logical) indicator if quantiles (including min and max) should be included in the table (default: TRUE).
incl_outliers: Per default we use fivenum to compute the quantiles (if quantiles = TRUE). If extreme values should be excluded from min/max in the table, boxplot( , plot = FALSE)$stats is used instead.
...: additional arguments. Currently not used.

Author

Benjamin Hofner

Examples

Run this code

if (require("nlme")) {
    ## Use dataset Orthodont
    data(Orthodont, package = "nlme")

    ## Get summary for continuous variables
    summarize(Orthodont, type = "numeric")

    ## Change statistics to display
    summarize(Orthodont, quantiles = FALSE, type = "numeric")
    summarize(Orthodont, quantiles = FALSE, count = FALSE, type = "numeric")
    summarize(Orthodont, mean_sd = FALSE, type = "numeric")

    ## for more examples see ?summarize
}