Note: For these exercises, we need the same packages as for the previous ones in this session: sjlabelled, tidyverse, haven.

Also, to make sure we’re all “on the same page” before we start the exercises, we should probably also import the data once more (and remove the labels).

allbus_2021 <- read_sav("./data/allbus_2021/ZA5280_v1-0-0.sav") %>% 
  remove_all_labels()

1

In this set of exercises, we want to work with the items from the ALLBUS 2021 that measure xenophobia (negative attitudes towards immigrants): ma01b, ma02, ma03, ma04. Using the across() function, the first thing we want to do with those variables is to convert the following values to NA: -42, -11, -10, -9.

The easiest (and shortest) option for doing this is to combine across() with a function from the sjlabelled package for specifying missing values.

solution 

allbus_2021 <- allbus_2021 %>%
  mutate(
    across(ma01b:ma04,
         ~set_na(
           .x,
           na = c(-42, -11, -10, -9))))

2

Now that we have properly defined the missing values for these variables, let’s create a new variable called xenophobia which is the mean score of these items.

The function we need to combine with across() for this is rowMeans(). Remember to specify that NAs should be ignored when the mean scores are computed (if you also want to include respondents that have not provided valid answers to all four questions).

solution 

allbus_2021 <- allbus_2021 %>%
  mutate(xenophobia = rowMeans(across(
    ma01b:ma04),
    na.rm = TRUE))