freq_function {EpiForsk}

Frequency Tables with Percentage and Odds Ratios

Description

A method for making 1- and 2-way frequency tables with percentages and odds ratios.

Usage

freq_function(
  normaldata,
  var1,
  var2 = NULL,
  by_vars = NULL,
  include_NA = FALSE,
  values_to_remove = NULL,
  weightvar = NULL,
  textvar = NULL,
  number_decimals = 2,
  output = c("all", "numeric", "col", "colw", "row", "roww", "total", "totalw"),
  chisquare = FALSE
)

Arguments

normaldata

A data frame or data frame extension (e.g. a tibble).

var1

A character string naming the first variable to get frequencies.

var2

An optional character naming the second variable to get frequencies. If NULL (standard) a 1-way frequency table of only var1 is created, and if var2 is specified a 2-way table is returned.

by_vars

An optional character vector naming variables in normal_data to stratify the calculations and output by. That is, ALL calculations will be made within the combinations of variables in the vector, hence it's possible to get N and % for many groups in one go.

include_NA

A logical. If FALSE (standard) missing variables (NA's) will be removed from var1 and var2. Any missing values in by_vars will not be removed. If TRUE all missing values will be included in calculations and the output.

values_to_remove

An optional character vector. When specified all values from var1 and var2 found in values_to_remove will be removed from the calculations and output.

weightvar

An optional character naming a column in normaldata with numeric weights for each observation. If NULL (standard) all observations have weight 1.

textvar

An optional character. When specified textvar is added to the resulting table as a comment. When NULL (standard) no such text addition is made.

number_decimals

A numeric indicating the number of decimals to show on percentages and weighted frequencies in the combined frequency and percent variables.

output

A character indicating the output type wanted:

• "all" - will give ALL output from tables. In many cases unnecessary and hard to get an overview of. This is set as the standard.

• "numeric" - will give frequencies and percents as numeric variables only, thus the number_decimals option is not in effect. This option might be useful when making figures/graphs.

• "col" - will only give unweighted number of observations and weighted column percent (if weights are used, otherwise unweighted)

• "colw" - will only give weighted number of observations and weighted column percent (if weights are used, otherwise unweighted)

• "row"- will only give unweighted number of observations and weighted row percent (if weights are used, otherwise unweighted). Only works in two-way tables (var2 is specified)

• "roww" - will only give weighted number of oberservations and weighted column percent (if weights are used, otherwise unweighted). Only works in two-way tables (var2 is specified)

• "total" - will only give unweighted number of observations and weighted percent of the total (if weights are used, otherwise unweighted). Only works in two-way tables (var2 is specified)

• "totalw" - will only give weighted number of observations and weighted percent of the total (if weights are used, otherwise unweighted). Only works in two-way tables (var2 is specified)

• Any other text will give the default ("all")

chisquare

A logical. FALSE (standard) will not calculate p-value for the chi-square test for two-way tables (var2 is specified). If TRUE, the table will include the chi-square p-value as well as the chi-square statistic and the corresponding degrees of freedom. It will be included in the output whichever output option have been specified. No chi-square test is performed or included in one-way tables (var2 is unspecified)

Value

A frequency table as a data frame object.

Author(s)

ASO

See Also

freq_function_repeated() to to get frequencies for multiple variables in one go.

Examples

data("starwars", package = "dplyr")

test_table1 <- freq_function(
  starwars,
  var1 = "homeworld"
)

test_table2 <- freq_function(
  starwars,
  var1 = "sex",
  var2 = "eye_color",
  output = "total"
)

test_table3 <- freq_function(
  starwars,
  var1 = "hair_color",
  var2 = "skin_color",
  by_vars = "gender",
  output = "col",
  number_decimals = 5
)