If we collapse these measurement variables to binary variables, we would get the following contingency table:

Weight	Height
	Short (<9 cm)	Tall (>9cm)	Total
Light (<36g)	3(a)	2 (b)	5
Heavy (>36g)	1 (c)	4 (d)	5
Totals	4	6	10

The risk ratio for the association of weight with height would then be 3/5 divided by 1/5 which gives 3.0. Since this ratio is larger than one it suggests there is a positive association between height and weight - it does not, however, suggest that weight causes height.

The following code produces a scatterplot using R.

weight=c(13, 17, 23, 27, 35, 38, 42, 45, 52, 57) height=c(3.2, 6.9, 10.2, 6.1, 10.3, 8.2, 12.4, 14.8, 10.2, 14.1) plot(weight, height, main='Scatterplot of height against weight', xlim=c(0,60), ylim=c(0,15))