Skip to contents

Convert a labelled vector into a factor

Source: R/make_factor.R
make_factor.Rd

make_factor() takes a labelled vector and converts it to a factor variable using the value labels. This works with numeric, character, and factor vectors.

This function is very similar to haven::as_factor() and sjlabelled::as_label() and is heavily based on both. However, it has some key differences. The main difference compared to both functions is that make_factor() adds a "transformation" attribute to the new variable indicating how it was created. You can see this in the examples.

Compared to sjlabelled::as_label() it is not as extensive. For example, while sjlabelled::as_label() works with data.frames and vectors, make_factor() only works with vectors. In addition, sjlabelled::as_label() has many different arguments that enable you to control the appearance of the labels, NAs, and other things. make_factor() on the other hand is much simpler. Similarly, haven::as_factor() also enables more customization over the output of the labels. Another key difference between this function and those is that if there are values without labels, this function returns an error.

Usage

make_factor(
  x,
  ordered = FALSE,
  drop_levels = TRUE,
  force = TRUE,
  na.rm = FALSE
)

Arguments

x

A vector with value labels. Can be numeric, character, or a factor

ordered

Logical. Determines if the factor be ordered. Defaults to TRUE.

drop_levels

Logical. Determines if unused factor levels should be dropped. Defaults to TRUE.

force

Logical. Determines if x should be forced to a vector even if there are no value labels. Defaults to TRUE.

na.rm

Logical. Determines if tags should be removed from NAs. Defaults to FALSE.

Value

A factor vector of same length as x.

Examples


library(adlgraphs)
library(dplyr)

# let's make a new variable and data set
new_df <- test_data %>%
  # convert top into a factor
  mutate(top_f = make_factor(top))

# compare the "top_f" to "top"
new_df %>% select(top, top_f)
#> # A tibble: 250 × 2
#>    top                   top_f            
#>    <dbl+lbl>             <fct>            
#>  1 1 [Strongly agree]    Strongly agree   
#>  2 2 [Somewhat agree]    Somewhat agree   
#>  3 2 [Somewhat agree]    Somewhat agree   
#>  4 3 [Somewhat disagree] Somewhat disagree
#>  5 2 [Somewhat agree]    Somewhat agree   
#>  6 4 [Strongly disagree] Strongly disagree
#>  7 2 [Somewhat agree]    Somewhat agree   
#>  8 2 [Somewhat agree]    Somewhat agree   
#>  9 2 [Somewhat agree]    Somewhat agree   
#> 10 4 [Strongly disagree] Strongly disagree
#> # ℹ 240 more rows

# check the attributes to see the label and transformation
attributes(new_df$top_f)
#> $levels
#> [1] "Strongly agree"    "Somewhat agree"    "Somewhat disagree"
#> [4] "Strongly disagree"
#> 
#> $class
#> [1] "factor"
#> 
#> $label
#> [1] "An ideal society requires some groups to be on top and others to be on the bottom"
#> 
#> $transformation
#> [1] "Converted 'top' into a factor based on its value labels"
#>