In brad-cannell/freqtables: Make Quick Descriptive Tables for Categorical Variables

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)

freqtables

The goal of freqtables is to quickly make tables of descriptive statistics for categorical variables (i.e., counts, percentages, confidence intervals). This package is designed to work in a tidyverse pipeline, and consideration has been given to get results from R to Microsoft Word ® with minimal pain.

Installation

You can install the released version of freqtables from CRAN with:

install.packages("freqtables")

And the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("brad-cannell/freqtables")

Example

Because freqtables is intended to be used in a dplyr pipeline, loading dplyr into your current R session is recommended.

library(dplyr)
library(freqtables)

The examples below will use R's built-in mtcars data set.

data("mtcars")

freq_table()

The freq_table() function produces one-way and two-way frequency tables for categorical variables. In addition to frequencies, the freq_table() function displays percentages, and the standard errors and confidence intervals of the percentages. For two-way tables only, freq_table() also displays row (subgroup) percentages, standard errors, and confidence intervals.

For one-way tables, the default 95 percent confidence intervals displayed are logit transformed confidence intervals equivalent to those used by Stata. Additionally, freq_table() will return Wald ("linear") confidence intervals if the argument to ci_type = "wald".

For two-way tables, freq_table() returns logit transformed confidence intervals equivalent to those used by Stata.

Here is an example of using freq_table() to create a one-way frequency table with all function arguments left at their default values:

mtcars %>% 
  freq_table(am)

Here is an example of using freq_table() to create a two-way frequency table with all function arguments left at their default values:

mtcars %>% 
  freq_table(am, cyl)

You can learn more about the freq_table() function and ways to adjust default behaviors in vignette("descriptive_analysis").

freq_test()

The freq_test() function is an S3 generic. It currently has methods for conducting hypothesis tests on one-way and two-way frequency tables. Further, it is made to work in a dplyr pipeline with the freq_table() function.

For the freq_table_two_way class, the methods used are Pearson's chi-square test of independence Fisher's exact test. When cell counts are <= 5, Fisher's Exact Test is considered more reliable.

Here is an example of using freq_test() to test the equality of proportions on a one-way frequency table with all function arguments left at their default values:

mtcars %>%
  freq_table(am) %>%
  freq_test() %>%
  select(var:percent, p_chi2_pearson)

Here is an example of using freq_test() to conduct a chi-square test of independence on a two-way frequency table with all function arguments left at their default values:

mtcars %>%
  freq_table(am, vs) %>%
  freq_test() %>%
  select(row_var:n, percent_row, p_chi2_pearson)

You can learn more about the freq_table() function and ways to adjust default behaviors in vignette("using_freq_test").

freq_format()

The freq_format function is intended to make it quick and easy to format the output of the freq_table function for tables that may be used for publication. For example, a proportion and 95% confidence interval could be formatted as "24.00 (21.00 - 27.00)."

mtcars %>%
  freq_table(am) %>%
  freq_format(
    recipe = "percent (lcl - ucl)",
    name = "percent_95",
    digits = 2
  ) %>%
  select(var, cat, percent_95)