1
$\begingroup$

I have two count data variables X and Y that contain many zero values (90% in X, 60% in Y). I would like to check if a correlation exists between these variables, but I'm not sure how to proceed due to presence of zero values.

Any suggestion?

$\endgroup$
1
$\begingroup$

I would try with a non-parametric correlation measure, like Spearman or Kendall correlations. This approach is more appropriate than using standard Pearson correlation.

$\endgroup$
  • $\begingroup$ When should I use Spearman or Kendall correlations? $\endgroup$ – user2314405 Apr 3 '14 at 13:54

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.