Question

You wish to examine if how much exercise a person does is related to sleep length. You think that people who exercise a lot sleep longer but your friend thinks that if someone exercises it will keep them awake. You measure 10 people and put them in order based on amount of exercise (1 = most exercise, 10 = least exercise) as well as sleep length (number of hours slept). What type of calculation would you perform (pearson r, spearman rho or sign test)? Explain your reasoning.

Answer 1

For Pearson's correlation coefficient, both the variables should be continouous.

Spearman's coefficient of association is computed when each of the variables is 'at least' ordinal.

Sign test is a non-parametric test for testing the significance of mean of one sample or equality of means of two samples.

The variable Sleep Length (in number of hours) is a continuous variable. The variable Amount of Exercise (on the scale of 1 to 10) is an ordinal variable. Therefore, Spearman's coefficient of association suits our case the most.

Hope this was helpful. Please leave back any comment.