Converting likert scale variables to be dichotomous

[lucy147]

I'm trying to perform a latent class analysis at the moment on a dataset. Most of the variables in the dataset are coded 1 to 10 on a likert scale e.g. 1 is least and 10 is most. I want to convert these variables so that 1:5 are grouped and labelled as 1 and then 6:10 are grouped and labelled as 2. I'm not sure if there is code that can do this?

I have tried recode(variable$data) but that does not seem to work.

Is there a command that I can use to recode the values?

[ihecker]

Hi! The case_when function of the dplyr package is probably the most efficient to recode variables.

[GreyMerchant]

As @ihecker mentioned, you can do something like the below:

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

x <- 1:10



dplyr::case_when(
  x <= 5 ~ 1,
  x >= 6 ~ 2
)
#>  [1] 1 1 1 1 1 2 2 2 2 2

^{Created on 2021-04-23 by the reprex package (v0.3.0)}

[lucy147]

Thank you! Will the variable stay coded like this for later analysis with the coded variable?

[GreyMerchant]

Here is an example of how you will retain it. You can call mutate which will then change or create a new column in your data frame. Here you see we create z based on column_a.

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

df <- data.frame(
"column_a" = 1:10,
"column_b" = 2:11
)

df <- df %>% 
  mutate(z =
dplyr::case_when(
  column_a <= 5 ~ 1,
  column_a >= 6 ~ 2
)
)

df
#>    column_a column_b z
#> 1         1        2 1
#> 2         2        3 1
#> 3         3        4 1
#> 4         4        5 1
#> 5         5        6 1
#> 6         6        7 2
#> 7         7        8 2
#> 8         8        9 2
#> 9         9       10 2
#> 10       10       11 2

^{Created on 2021-04-23 by the reprex package (v0.3.0)}

[lucy147]

Thanks again, is there any way to add to this code to make values >10 appear as NA?

[GreyMerchant]

Here you go

library(dplyr)
#> Warning: package 'dplyr' was built under R version 4.0.5
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

df <- data.frame(
  "column_a" = 1:12,
  "column_b" = 2:13
)

df <- df %>% 
  mutate(z =
           dplyr::case_when(
             column_a <= 5 ~ 1,
             column_a >= 6 & column_a <= 10  ~ 2,
             column_a > 10 ~ NA_real_
           )
  )


df
#>    column_a column_b  z
#> 1         1        2  1
#> 2         2        3  1
#> 3         3        4  1
#> 4         4        5  1
#> 5         5        6  1
#> 6         6        7  2
#> 7         7        8  2
#> 8         8        9  2
#> 9         9       10  2
#> 10       10       11  2
#> 11       11       12 NA
#> 12       12       13 NA

^{Created on 2021-04-23 by the reprex package (v0.3.0)}