In this final set of exercises for the data wrangling basics, we will transform and recode variables, and work with missing data. In the following exercises, we will focus on data wrangling functions from the tidyverse.

Same procedure as before: Load the tidyverse package(s) and import the data and have the codebook ready.

library(tidyverse)

gp_covid <- read_csv2("./data/ZA5667_v1-1-0.csv")

1

To begin with, let’s use the dplyrfunction for creating and transforming variables to create a new variable representing political orientation named pol_orientation_new that ranges from 1 to 11 instead of from 0 to 10 as is the case for the original variable political_orientation.

We simply need to add 1 to the existing variable.

solution 

gp_covid <- gp_covid %>% 
  mutate(pol_orientation_new = political_orientation + 1)

2

By combining the function used for the previous task with another one from the dplyr package, recode the values of the variable measuring trust in the federal government with regard to dealing with the Corona virus into a new variable named distrust_gov that captures distrust instead of trust.

The name of the variable we want to transform is hzcy048a. Disregarding missing values for the moment, its values range from 1 to 5. Remember that the correct syntax for recoding values with the corresponding dplyr function is old value (enclosed in backticks) = new value.

solution 

gp_covid <- gp_covid %>% 
  mutate(distrust_gov = recode(hzcy048a,
                               `5` = 1, # old_value = new_value
                               `4` = 2,
                               `2` = 4,
                               `1` = 5))

3

The variable we have just recoded still contains several values representing different types of missing values. Using the appropriate dplyr function, recode the following values as NA for the new distrust_gov variable: -99, -77, -33, and 98.

To to this, we need to combine mutate() with the dplyr function for recoding specific values as NA.

solution 

gp_covid <- gp_covid %>% 
  mutate(distrust_gov = na_if(distrust_gov, -99)) %>% 
  mutate(distrust_gov = na_if(distrust_gov, -77)) %>% 
  mutate(distrust_gov = na_if(distrust_gov, -33)) %>%
  mutate(distrust_gov = na_if(distrust_gov, 98))

4

After recoding a set of values as NA for one variable, let’s now do the same for the whole gp_covid data frame. This time, however, we do not want to recode 98 as NA as it is a valid value for the id variable.

This time, we do not need the mutate() function.

solution 

gp_covid <- gp_covid %>% 
  na_if(-99) %>% 
  na_if(-77) %>% 
  na_if(-33)

5

As na_if() only takes only takes single values as its second argument (i.e., the value to replace with NA), let’s use a function from the sjlabelled function to achieve the same thing with fewer lines of code.

The function we are looking for can also be included in a pipe chain and takes a vector of values to be recoded as NA as its second (required) argument.

solution 

library(sjlabelled)

gp_covid <- gp_covid %>% 
  set_na(na = c(-99, -77, -33))

6

How many of the respondents do not have a missing value for the variable political_orientation? To answer this question, please use a function from the tidyr package that allows you to exclude cases with missing values. Do not assign the result to a new object.

To count the number of cases, you can use the base R function nrow() at the end of your pipe.

solution 

gp_covid %>% 
  drop_na(political_orientation) %>% 
  nrow()
## [1] 3678

7

As a final exercise for this session, let’s recode the marstat variable into an unordered factor called marstat_fac that has 4 levels named after the different value labels listed in the codebook.

The value labels from the codebook are 1 = Married, 2 = Single, 3 = Divorced, 4 = Widowed. The dplyr function we need to use here (in combination with mutate()) is recode_factor().

solution 

gp_covid <- gp_covid %>% 
  mutate(marstat_fac = recode_factor(marstat,
                                     `1` = "Married",
                                     `2` = "Single",
                                     `3` = "Divorced",
                                     `4` = "Widowed"))