I've looked at a few of these types of posts but can't seem to find a concrete answer. I am investigating the relationship between 'risk aversion' and 'human capital' for a project.

Risk aversion: Acceptable risk is used as a measurement and it takes a value between 10% and 100%, in increments of 10%. (I.E. 10 possible values where 10% = risk taker and 100% = risk averse). [DISCRETE]

Human capital: Represents the present value of all future earnings and depends on a number of factors including age, salary and occupation. [CONTINUOUS]

I performed a Pearson correlation between these two variables without really looking into the specifics too much. It finds a small negative correlation which is as expected.

However, I have since then read that the Pearson correlation is only designed to measure the correlation between two continuous variables and so I don't think I can use this result. Am I correct in saying this and if so how can I test this relationship (including significance).


marked as duplicate by kjetil b halvorsen, Michael Chernick, gung, Peter Flom Apr 17 '17 at 13:10

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • $\begingroup$ Yes the Pearson correlation coefficient does compare the relationship between two continuous variables and hence is not appropriate for your situation. You can look at other methods of association such as Spearman's rho or Kendall's tau. $\endgroup$ – Michael Chernick Apr 17 '17 at 11:57

It would be better to use Spearman's rank correlation coefficient, which makes less assumptions on the variables than Pearson's coefficient, thus allowing the calculation of the association between the tanks of two variables which are either ordinal or continuous. According to Wikipedia, the Spearman coefficient is "appropriate for both continuous and discrete variables, including ordinal variables".


Not the answer you're looking for? Browse other questions tagged or ask your own question.