I work with wind data. In order to estimate the strength of the correlation between two anenometers, in my department they usually calculate the pearson R.

As far as I know the pearson R should be used under 2 assumptions: -Linear relationship between the variables (yes) -The data follows a normal distribution (no, as wind data has a weibull distribution)

Should we use then a non parametric coefficient like spearman to calculate the correlation?


1 Answer 1


Pearson correlation does not assume that the data are normally distributed but the Pearson and Spearman correlations ask different questions. Pearson lets outliers have a lot of influence, Spearman does not. Which you want is up to you.

Exact results will depend on the shape parameter, but Weibull has long tails, so it has outliers.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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