Skip to contents

Check covariate missingness and imputed-version coverage

Source: R/check_study_data.R
check_missingness_nona.Rd

For each base covariate, reports the number and percentage of missing values, whether an imputed companion column exists (identified by nona_suffix), and whether that companion itself contains any missing values.

Usage

check_missingness_nona(
  data,
  study_id = NULL,
  covariates = NULL,
  prefix = "X_",
  nona_suffix = "_nona",
  exclude = NULL
)

Arguments

data

A data frame or tibble.

study_id

Optional character scalar. If provided, a study_id column is appended to the returned tibble.

covariates

Columns to check. Supply unquoted names or tidyselect helpers (e.g., starts_with("cov_")), or a character vector of column names. If omitted, all columns starting with prefix are used (excluding nona_suffix and "_missing" suffixes).

prefix

Character string. Prefix used to auto-select covariate columns when covariates is omitted (default: "X_").

nona_suffix

Character string. Suffix that identifies imputed companion columns (default: "_nona"). Used both to exclude companion columns from the base selection and to look them up when checking nona_has_na.

exclude

Additional columns to drop from the selection. Supply unquoted names or tidyselect helpers (e.g., ends_with("_old")), or a character vector of exact column names. Applied after covariates is resolved. NULL (the default) means no additional exclusions.

Value

A tibble with columns variable, n_missing, pct_missing, has_nona_version, nona_has_na, and optionally study_id. nona_has_na is NA when no companion column exists.

Details

When covariates is omitted, all columns whose names start with prefix are selected, excluding any that end with nona_suffix or "_missing". Supply covariates to override this default, or exclude to drop additional columns from whatever was selected.

See also

Other per-study checks: check_attrition(), check_balance(), check_covariate_missingness(), check_smd(), check_y_bounds()

Examples

dat <- data.frame(
  X_age      = c(25, NA, 30),
  X_age_nona = c(25, 27, 30),
  X_income   = c(NA, 50, 60)
)
check_missingness_nona(dat)
#> # A tibble: 2 × 5
#>   variable n_missing pct_missing has_nona_version nona_has_na
#>   <chr>        <int>       <dbl> <lgl>            <lgl>      
#> 1 X_age            1       0.333 TRUE             FALSE      
#> 2 X_income         1       0.333 FALSE            NA         
check_missingness_nona(dat, study_id = "my_study")
#> # A tibble: 2 × 6
#>   variable n_missing pct_missing has_nona_version nona_has_na study_id
#>   <chr>        <int>       <dbl> <lgl>            <lgl>       <chr>   
#> 1 X_age            1       0.333 TRUE             FALSE       my_study
#> 2 X_income         1       0.333 FALSE            NA          my_study
check_missingness_nona(dat, covariates = "X_age")
#> # A tibble: 1 × 5
#>   variable n_missing pct_missing has_nona_version nona_has_na
#>   <chr>        <int>       <dbl> <lgl>            <lgl>      
#> 1 X_age            1       0.333 TRUE             FALSE      
check_missingness_nona(dat, nona_suffix = "_imputed")
#> # A tibble: 3 × 5
#>   variable   n_missing pct_missing has_nona_version nona_has_na
#>   <chr>          <int>       <dbl> <lgl>            <lgl>      
#> 1 X_age              1       0.333 FALSE            NA         
#> 2 X_age_nona         0       0     FALSE            NA         
#> 3 X_income           1       0.333 FALSE            NA