1
$\begingroup$

How do i check that two random variables satisfy the requirements for computing correlation via Pearson or Spearman? If the data is non-linear/ non-Monotonic, are there other tests i could use to compute a correlation coefficient?

$\endgroup$
2
  • $\begingroup$ Once the variables are random, there is nothing like linear and non-linear. Only thing you need is adequate/ minimum sample size to cope with population. $\endgroup$ Sep 12 '14 at 15:30
  • $\begingroup$ the tests are not needed. What is needed is to examine data-type. $\endgroup$ Sep 12 '14 at 15:35
0
$\begingroup$

Distance correlation takes also nonlinear correlations into account.

$\endgroup$
1
  • $\begingroup$ The link in your answer is not hyperlinking correctly. Please try putting triangle braces (< >) around it or putting space between the parenthesis and link. $\endgroup$
    – user41994
    Feb 23 '13 at 12:33
0
$\begingroup$

Correlation is a measure of association between two variables. It does not establish linearity or no linearity. It provides us with indication whether the relationship is generally positive or negative and its average magnitude.

If one has grouped data on each of two variables, Pearson correlation may be calculated. And also if we have continuous data, the Karl Pearson formula of correlation coefficient can be applied. Spearman correlation requires the rank database.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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