Replace missing value placeholders with NA

Usage

recode_missing(df, dict_df, var)

Arguments

df

A data frame containing raw survey data. Must have columns entity_id and the variable specified by var.

dict_df

A data frame. Data dictionary for df from the data provider. Expected columns:

variable: variable name
name: value code
missing: whether the value is a missing value placeholder
label: value label

var

A string indicating the name of the variable in df to clean.

Value

A data frame with the same number of rows as df with the following columns: entity_id and the cleaned variable, with missing value placeholders replaced by NA.

Details

Missing codes are replaced with a plain NA, regardless of the reason for missingness. Specialized NA values (e.g., haven's NA(a) or NA(z)) are not used, so distinctions such as "refused" vs. "don't know" are not preserved.

Examples

df <- data.frame(
  entity_id = c(8291, 8657, 3279, 9743, 5952),
  GEN_HLTH_COM = c(1, 4, 5, 3, 9)
)

dict_df <- data.frame(
  variable = rep("GEN_HLTH_COM", 7),
  name = c(1:5, 8, 9),
  missing = c(rep(0, 5), rep(1, 2)),
  label = c(
    "Excellent",
    "Very good",
    "Good",
    "Fair",
    "Poor",
    "[DO NOT READ] Don't know/No answer",
    "[DO NOT READ] Refused"
  )
)

recode_missing(df, dict_df, "GEN_HLTH_COM")
#>   entity_id GEN_HLTH_COM
#> 1      8291            1
#> 2      8657            4
#> 3      3279            5
#> 4      9743            3
#> 5      5952           NA