Skip to contents

Descriptive Statistics for Univariate Data with Optional Reference Comparison

Source: R/s_summary_diff.R
s_summary_diff.Rd

[Experimental]

Computes descriptive statistics for a single variable df[[.var]] using tern::s_summary(), which dispatches type-specific methods depending on the S3 class of the input (e.g., character, factor, logical, numeric).

Optionally, it computes estimates and associated inferential statistics for the difference in population means between two underlying distributions, based on observed samples df[[.var]] and .ref_group[[.var]] using s_diff_means(). These statistics are applicable only to numeric data.

Usage

s_summary_diff(
  df,
  .var,
  .stats = NULL,
  .ref_group = NULL,
  .in_ref_col = FALSE,
  control = tern::control_analyze_vars(),
  ...
)

Arguments

df

(data.frame)
data set containing all analysis variables.

.var

(character(1))
Name of the column in df containing the values for which statistics are computed. The variable type is handled by the corresponding methods of tern::s_summary(). If difference-in-means statistics are requested, df[[.var]] must be numeric.

.stats

(character or NULL)
Names of statistics to be computed. For numerical data, the names of supported statistics are listed via tern::get_stats() for statistics for a single variable and junco_get_stats("diff_means") for difference-in-means statistics. If NULL, all available statistics for numerical data are computed.

.ref_group

(data.frame or vector)
the data corresponding to the reference group.

.in_ref_col

(logical)
TRUE when working with the reference level, FALSE otherwise.

control

(list)
List of control options passed to tern::s_summary(). If difference-in-means statistics are requested, control$conf_level specifies the confidence level of the interval.

...

Additional named arguments passed to s_diff_means(), except df1, df2, .var, and conf.level, which are already passed.

Value

A named list with the requested statistics.

Author

WW

Examples

df <- data.frame(
  USUBJID = c("X01", "X02", "X03", "X04", "X05"),
  TRT01A = rep("ARM_A", 5),
  PARAMCD = rep("SYSBP", 5),
  AVISIT = rep("Visit 1", 5),
  CHG = c(4, 1, -1, 9, -2)
)
df
#>   USUBJID TRT01A PARAMCD  AVISIT CHG
#> 1     X01  ARM_A   SYSBP Visit 1   4
#> 2     X02  ARM_A   SYSBP Visit 1   1
#> 3     X03  ARM_A   SYSBP Visit 1  -1
#> 4     X04  ARM_A   SYSBP Visit 1   9
#> 5     X05  ARM_A   SYSBP Visit 1  -2

rg <- data.frame(
  USUBJID = c("X06", "X07", "X08", "X09", "X10"),
  TRT01A = rep("Placebo", 5),
  PARAMCD = rep("SYSBP", 5),
  AVISIT = rep("Visit 1", 5),
  CHG = c(-2, 6, -2, 5, 2)
)
rg
#>   USUBJID  TRT01A PARAMCD  AVISIT CHG
#> 1     X06 Placebo   SYSBP Visit 1  -2
#> 2     X07 Placebo   SYSBP Visit 1   6
#> 3     X08 Placebo   SYSBP Visit 1  -2
#> 4     X09 Placebo   SYSBP Visit 1   5
#> 5     X10 Placebo   SYSBP Visit 1   2

.stats <- c("n", "mean_sd", "diff_means_n1", "diff_means_n2", "diff_means_est_ci")

# With reference group.
s_summary_diff(df, "CHG", .stats, rg)
#> $n
#> n 
#> 5 
#> 
#> $mean_sd
#>     mean       sd 
#> 2.200000 4.438468 
#> 
#> $diff_means_n1
#> diff_means_n1 
#>             5 
#> attr(,"label")
#> [1] "Difference in Means Sample Size (Group 1)"
#> 
#> $diff_means_n2
#> diff_means_n2 
#>             5 
#> attr(,"label")
#> [1] "Difference in Means Sample Size (Group 2)"
#> 
#> $diff_means_est_ci
#>    diff_means_est diff_means_ci_lwr diff_means_ci_upr 
#>          0.400000         -5.632096          6.432096 
#> attr(,"conf.level")
#> [1] 0.95
#> attr(,"label")
#> [1] "Difference in Means (95% CI)"
#> 

# Using df as reference.
s_summary_diff(df, "CHG", .stats, df, .in_ref_col = TRUE)
#> $n
#> n 
#> 5 
#> 
#> $mean_sd
#>     mean       sd 
#> 2.200000 4.438468 
#> 
#> $diff_means_n1
#> NULL
#> 
#> $diff_means_n2
#> NULL
#> 
#> $diff_means_est_ci
#> NULL
#>