Label factor levels from a lookup table (using labelled package)

Up front, I'm not a labelled-guru, but on reading https://larmarange.github.io/labelled/articles/labelled.html#value-labels, one way is to apply labelled() to each column of dummydat with the applicable element of dictest.

for (nm in names(dummydat)) dummydat[[nm]] % in your code, that suggests you're using dplyr and could likely use tidyr, so I'll offer a more direct path for your code snippets:
library(dplyr)
library(tidyr)
library(labelled)
dictest 
  mutate(name = gsub("x", "", name)) |>
  dictionarytovaluelabels(namesfrom = question, valuesfrom = name, labelsfrom = value, data = dummydat)
dummydat2 
  mutate(across(everything(), ~ labelled(.x, dictest[[curcolumn()]])))
dummydat2
  X1 X21 X22 X3
1  1    1    3  1
2  2    2    2  2
3  1    3    1  3
dummydat2$X22

[1] 3 2 1
Labels:
 value     label
     1     never
     2 sometimes
     3    always

dictionarytovaluelabels() returns a named list of label definitions; splice it into setvaluelabels() with !!! to apply all labels to the data at once.

library(labelled)
dummydatlab 
  transform(variable = as.numeric(gsub("x", "", variable))) |>
  na.omit() |>
  dictionarytovaluelabels(
    namesfrom = question,
    valuesfrom = variable,
    labelsfrom = value,
    data = dummydat
  ) |>
  (\(dic) setvaluelabels(dummydat, !!!dic))()
str(dummydatlab)
#> 'data.frame':    3 obs. of  4 variables:
#>  $ X1  : dbl+lbl [1:3] 1, 2, 1
#>    ..@ labels: Named num  1 2
#>    .. ..- attr(*, "names")= chr [1:2] "yes" "no"
#>  $ X21: dbl+lbl [1:3] 1, 2, 3
#>    ..@ labels: Named num  1 2 3
#>    .. ..- attr(, "names")= chr [1:3] "never" "sometimes" "always"
#>  $ X2_2: dbl+lbl [1:3] 3, 2, 1
#>    ..@ labels: Named num  1 2 3
#>    .. ..- attr(, "names")= chr [1:3] "never" "sometimes" "always"
#>  $ X3  : num  1 2 3

^{Created on 2026-02-27 with reprex v2.1.1}