18 Paired Numerical Samples

We work with the famous dataset that Pearson collected on the heights of men and their sons.

18.1 Scatterplot

This is an appropriate plot for paired numerical data.

require(UsingR)
plot(father.son, pch = 16, xlab = "father's height", ylab = "son's height", asp = 1)
abline(1, 0, lty = 2) # diagonal line (congruent with what follows)

18.2 Testing for symmetry

The observations are paired (father, son). We take difference and test for symmetry using the Wilcoxon signed-rank test. (According to the manual, the exact p-value, i.e., the permutation p-value, is returned when there are 50 observations or fewer.)

wilcox.test(father.son$sheight, father.son$fheight, paired = TRUE)

    Wilcoxon signed rank test with continuity correction

data:  father.son$sheight and father.son$fheight
V = 405955, p-value < 2.2e-16
alternative hypothesis: true location shift is not equal to 0

There is strong evidence that the distribution of the difference in heights is not symmetric (about 0). This is apparent when plotting a histogram.

hist(father.son$sheight - father.son$fheight, breaks = 50, col = "grey", xlab = "difference in height")
abline(v = 0, lty = 2, lwd = 2)