Should we be using factors instead of character vectors?

[rpsychologist]

Wouldn't it be better if all categorical variables were factors? For example these (and probably others):

lab-materials/labs/probability-rules.qmd

Lines 27 to 29 in 09b85f8

	d_bl |>
	select(where(is.character)) |>
	glimpse()

Rows: 181
Columns: 5
$ gender    <chr> "Woman", "Man", "Man", "Man", "Woman", "Woman", "Woman", "Ma…
$ education <chr> "Primary", "University", "Secondary", "Secondary", "Universi…
$ income    <chr> "Medium", "Medium", "Medium", "Medium", "Medium", "Low", "Lo…
$ gad_cat   <chr> "Low anxiety", "High anxiety", "High anxiety", "Low anxiety"…
$ phq_cat   <chr> "Low depression", "High depression", "High depression", "Low…