num_stat: Compute some basic descriptive statistics.

In shug0131/cctu: In-House Functions for Report Writing at Cambridge Clinical Trials Unit

Compute some basic descriptive statistics.

Description

Values of type factor, character and logical are treated as categorical. For logicals, the two categories are given the labels 'Yes' for TRUE, and 'No' for FALSE. Factor levels with zero counts are retained.

Usage

num_stat(x, digits = 3, digits_pct = 1, rounding_fn = signif_pad)

cat_stat(x, digits_pct = 1)

Arguments

x	A vector or numeric, factor, character or logical values.
digits	An integer specifying the number of significant digits to keep for numerical results. See signif_pad.
digits_pct	An integer specifying the number of significant digits to keep for percentage.
rounding_fn	The function to use to do the rounding. Defaults to signif_pad.

Value

A list. For numeric x, the list contains the numeric elements:

• N: the number of non-missing values

• NMISS: the number of missing values

• SUM: the sum of the non-missing values

• MEAN: the mean of the non-missing values

• SD: the standard deviation of the non-missing values

• MIN: the minimum of the non-missing values

• MEDIAN: the median of the non-missing values

• CV: the percent coefficient of variation of the non-missing values

• GMEAN: the geometric mean of the non-missing values if non-negative, or NA

• GCV: the percent geometric coefficient of variation of the non-missing values if non-negative, or NA

• GSD: the geometric standard deviation of the non-missing values if non-negative, or NA

• Q1: the first quartile of the non-missing values (alias q25)

• Q2: the second quartile of the non-missing values (alias q50 or Median)

• Q3: the third quartile of the non-missing values (alias q75)

• IQR: the inter-quartile range of the non-missing values (i.e., Q3 - Q1)

If x is categorical (i.e. factor, character or logical), the list contains a sublist for each category, where each sublist contains the numeric elements:

• FREQ: the frequency count

• PCT: the percent relative frequency, including NA in the denominator

• PCTnoNA: the percent relative frequency, excluding NA from the denominator

• Nall: total count, including NA from the denominator

• N: total count, excluding NA from the denominator

See Also

signif_pad round_pad

Examples

x <- exp(rnorm(100, 1, 1))
num_stat(x)

y <- factor(sample(0:1, 99, replace = TRUE), labels = c("Female", "Male"))
y[1:10] <- NA
cat_stat(y)
cat_stat(is.na(y))

shug0131/cctu documentation built on Feb. 15, 2025, 3:27 p.m.