Skip to contents

Compute a cross-tabulation of two categorical variables

Source: R/utils.R
compute_crosstab.Rd

Produces a wide-format contingency table of row_var (rows) by col_var (columns), including row and column totals. When a strat_var is supplied the table is computed separately for each level of the stratification variable and the results are stacked with a leading Stratum column.

Usage

compute_crosstab(df, row_var, col_var, strat_var = NULL)

Arguments

df

A data.frame or tibble.

row_var

Character string. Name of the row variable (e.g. "SEX").

col_var

Character string. Name of the column variable (e.g. "RACE").

strat_var

Character string or NULL. Optional stratification variable (e.g. "TRT01P"). Pass NULL or "" for an unstratified table.

Value

A data.frame in wide format:

  • Column 1 (or 2 if stratified): row_var levels plus a "Total" row.

  • Middle columns: one column per col_var level.

  • Last column: Total (row sums).

  • If strat_var is given, a leading Stratum column identifies each stratum. A grand-total block across all strata is not appended automatically — compute the unstratified table for that.

Details

Missing values in any of the three variables are displayed as "(Missing)" rather than being silently dropped, so analysts can spot incomplete records.

Examples

df <- data.frame(
  SEX  = c("M","F","M","F","M","F"),
  RACE = c("White","White","Black","Asian","Black","White"),
  TRT  = c("A","A","B","B","A","B")
)
compute_crosstab(df, "SEX", "RACE")
#>     SEX Asian Black White Total
#> 1     F     1     0     2     3
#> 2     M     0     2     1     3
#> 3 Total     1     2     3     6
compute_crosstab(df, "SEX", "RACE", strat_var = "TRT")
#>   Stratum   SEX Black White Total Asian
#> 1 TRT = A     F     0     1     1    NA
#> 2 TRT = A     M     1     1     2    NA
#> 3 TRT = A Total     1     2     3    NA
#> 4 TRT = B     F     0     1     2     1
#> 5 TRT = B     M     1     0     1     0
#> 6 TRT = B Total     1     1     3     1