Label attributes with a few renames

Builds a value for the labels argument of the descriptive tables (table_categorical(), table_continuous(), table_continuous_lm(), table_outcome(), table_categorical_svy() and table_continuous_svy()) and of the regression tables (table_regression() and table_regression_uv()). The variables it names take the labels given here, and every other variable shows its label attribute (as set by haven, labelled or label_from_names()), or its name when it has none, whatever options(spicy.var_labels) says. var_labels() with no argument is the same as labels = TRUE.

A plain named vector, labels = c(bmi = "Body mass index"), differs in one point: the variables it leaves out follow options(spicy.var_labels), which shows their names unless it is TRUE.

In a regression table, the names are the model's: the outcome, the term labels (age, or sex for the header of a factor), and the coefficient names (sexMale, (Intercept)). The attribute names the outcome, the numeric terms, and the factor headers. Factor levels, interactions, and transformed terms keep their names.

Usage

var_labels(...)

# S3 method for class 'spicy_var_labels'
print(x, ...)

Arguments

...

name = "label" pairs, one per variable to rename, each label a single string. A named character vector can be spliced with !!!. Names that match nothing in a table are ignored. For print(), passed on to print().

x

A spicy_var_labels object, for print().

Value

A named character vector of class spicy_var_labels, which prints as the plain named vector. print() returns x invisibly.

See also

The labels argument of table_categorical() and its siblings, and of table_regression(), which lists the four forms it takes.

Examples

# The label attribute of every column, one rename.
table_continuous(
  sochealth,
  select = c(bmi, age),
  labels = var_labels(bmi = "BMI (kg/m2)")
)
#> Descriptive statistics
#> 
#>  Variable    │   M     SD     Min    Max   95% CI LL  95% CI UL   n   
#> ─────────────┼────────────────────────────────────────────────────────
#>  BMI (kg/m2) │ 25.93   3.72  16.00  38.90    25.72      26.14    1188 
#>  Age (years) │ 49.26  14.70  25.00  75.00    48.43      50.10    1200 
#> 
#> Note. Missing values removed: BMI (kg/m2) (12). CI = confidence interval; LL = lower limit; UL = upper limit.

# A vector of renames kept for a whole project, spliced.
renames <- c(bmi = "BMI (kg/m2)", smoking = "Smoker")
tbl <- table_categorical(
  sochealth,
  select = smoking,
  by = sex,
  labels = var_labels(!!!renames)
)

# A regression table: the outcome and the factor header keep their
# label attribute, the numeric term takes the one given here.
table_regression(
  lm(wellbeing_score ~ age + sex, data = sochealth),
  labels = var_labels(age = "Age in years")
)
#> Linear regression: WHO-5 wellbeing index (0-100)
#> 
#>  Variable        │    B      SE       95% CI        p   
#> ─────────────────┼──────────────────────────────────────
#>  (Intercept)     │   65.07  1.63  [61.87, 68.26]  <.001 
#>  Age in years    │    0.04  0.03  [-0.02,  0.10]   .163 
#>  Sex:            │                                      
#>    Female (ref.) │     –     –          –          –    
#>    Male          │    3.90  0.90  [ 2.14,  5.65]  <.001 
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#>  n               │ 1200                                 
#>  R²              │    0.02                              
#>  Adj. R²         │    0.02                              
#> 
#> Note. Linear regression. Std. errors: classical (OLS). CI = confidence interval.