Quick Univariate Summary

 

Menu locations:
Analysis_Descriptive_Quick Summary
Edit_Describe Column Data (or press F9 in the worksheet)

 

This function provides rapid access to descriptive statistics for a worksheet column of data.

 

Shortcut: click on the right mouse button when the mouse cursor is over the column of data you want to describe and you will be given summary statistics for that column, provided the setting of the Edit_Options menu item is set to "Column summary".

 

The statistics calculated here are a sub-set of those available through the Analysis_Descriptive_Descriptive Report menu function. If you want to calculate summary statistics for more than one column at a time then you must use the Analysis_Descriptive_Descriptive Report menu function.

 

For definitions of the statistics calculated, please see descriptive report.

 

Example

Test workbook (Parametric worksheet: Systolic BP).

 

Consider the 20 resting systolic blood pressures of the single sample t test example. Open the test workbook, click on the column marked "Systolic BP" and press F9, or select Quick Univariate Summary from the Descriptive section of the Analysis menu:

 

Title: Systolic BP  
Valid data 20
Missing 0
Sum 2601
Mean 130.05
Variance 99.207895
Standard deviation 9.960316
Variation coefficient 0.076588
Standard error of mean 2.227194
95% Upper CL of mean 134.711571
95% Lower CL of mean 125.388429
Geometric mean 129.684825
Skewness -0.047857
Kurtosis 2.411345
Maximum 149
95th percentile 146.5
Upper quartile 137.5
Median 130
Lower quartile 124
Interquartile range 13.5
5th percentile 113
Minimum 111
Range 38

 

The quick summary always uses centile type 1 for the quartiles and percentiles, and shows six decimal places whatever the setting in Analysis_Options.

 

This R code reproduces the example above. It needs no packages and was checked with R 4.6.1. Paste it into R, or save it as a script and run it.

# Quick univariate summary: the StatsDirect help example (resting systolic blood
# pressure of 20 first year resident doctors, from the single sample t test) in R
bp <- c(128, 127, 118, 115, 144, 142, 133, 140, 132, 131, 111, 132, 149, 122, 139,
        119, 136, 129, 126, 128)
n <- length(bp)

# R's standard summary, then each line of the quick summary to 6 places
print(summary(bp))
six <- function(x) formatC(round(x, 6), digits = 6, format = "f", drop0trailing = TRUE)
cat("Valid data", n, "  Missing", sum(is.na(bp)), "\n")
cat("Sum", six(sum(bp)), "\n")
cat("Mean", six(mean(bp)), "\n")
cat("Variance", six(var(bp)), "\n")
cat("Standard deviation", six(sd(bp)), "\n")
cat("Variation coefficient", six(sd(bp) / mean(bp)), "\n")
se <- sd(bp) / sqrt(n)
cat("Standard error of mean", six(se), "\n")
cat("95% Upper CL of mean", six(mean(bp) + qt(0.975, n - 1) * se), "\n")
cat("95% Lower CL of mean", six(mean(bp) - qt(0.975, n - 1) * se), "\n")
cat("Geometric mean", six(exp(mean(log(bp)))), "\n")

# Skewness and kurtosis are sqrt(b1) and b2, from the moments about the mean
# with n as the divisor
m <- function(k) mean((bp - mean(bp))^k)
cat("Skewness", six(m(3) / m(2)^1.5), "\n")
cat("Kurtosis", six(m(4) / m(2)^2), "\n")

# The quick summary uses centile type 1: the ordered value whose cumulative
# count first exceeds pn, or the average of two values when pn is a whole
# number, which is type 2 in R
q <- quantile(bp, c(0.05, 0.25, 0.5, 0.75, 0.95), type = 2)
cat("Maximum", six(max(bp)), "\n")
cat("95th percentile", six(q[5]), "\n")
cat("Upper quartile", six(q[4]), "\n")
cat("Median", six(q[3]), "\n")
cat("Lower quartile", six(q[2]), "\n")
cat("Interquartile range", six(q[4] - q[2]), "\n")
cat("5th percentile", six(q[1]), "\n")
cat("Minimum", six(min(bp)), "\n")
cat("Range", six(max(bp) - min(bp)), "\n")