describeBy: Descriptive statistics

In AlineTalhouk/Amisc: Miscellaneous Utility Functions

Descriptive statistics

Description

Descriptive statistics and univariable association tests

Usage

describeBy(
  data,
  var.names,
  var.labels = var.names,
  by1,
  by2 = NULL,
  total = c("top", "bottom", "none"),
  Missing = TRUE,
  test = TRUE,
  digits = 0,
  total.digits = 0,
  p.digits = 3,
  bold_pval = FALSE,
  sig.level = 0.05,
  dispersion = c("sd", "se"),
  stats = c("parametric", "non-parametric"),
  per = "col",
  simulate.p.value = FALSE,
  B = 2000,
  bold_var = TRUE,
  fill = FALSE
)

Arguments

data	data.frame to produce descriptive statistics
var.names	variable names of interest in data
var.labels	variable descriptions. Uses var.names by default.
by1	factor to split other variables by in data
by2	optional second factor to split other variables by
total	add a row showing the total counts of each by1 level at the top or bottom of the table. Setting none hides the total row.
Missing	logical; if TRUE, shows missing value counts, if they exist
test	logical; if TRUE, univariable tests are performed and a PValue column is added to the end of the table.
digits	number of digits to round descriptive statistics. Supply a single value to round all variables to the same number of digits, or a vector of values to supply different rounding per variable.
total.digits	number of digits to round the total count percentages
p.digits	number of digits to round univariable test p-value
bold_pval	logical; if TRUE, p-values are bolded if statistically significant at sig.level
sig.level	significance level; default 0.05
dispersion	measure of variability, either "sd" (default) or "se".
stats	either "parametric" (default) or "non-parametric" univariable tests are performed for continuous variables. We use the parametric one-way test or the non-parametric Kruskal-Wallis test.
per	print column ("col") or row ("row") percentages. Suppress percentages with "none".
simulate.p.value	passed to chisq.test. Only relevant for categorical variables.
B	passed to chisq.test. Only relevant for categorical variables.
bold_var	logical; if TRUE, the Variable names are wrapped in double asterisks. If the table is parsed by pandoc the variable names are in bold.
fill	logical; if TRUE, the Variable and PValue columns are repeated for every row it pertains to. If FALSE, the value is only shown when it changes.

Details

Takes variables from data and returns descriptive statistics split on factor by1.

Value

A table with descriptive statistics for continuous and categorical variables and relevant univariable association tests

Author(s)

Aline Talhouk

Examples

mtcars$cyl <- as.factor(mtcars$cyl)
mtcars$vs <- as.character(mtcars$vs)
Amisc::describeBy(data = mtcars, var.names = c("vs", "hp"), by1 = "cyl",
Missing = TRUE, dispersion = "sd", stats = "parametric")

AlineTalhouk/Amisc documentation built on May 26, 2023, 3:40 p.m.