A quick question on Spearman's rank correlation coefficient

[Just some guy]

Hi,

I was just wondering about spearman's rank correlation coefficient hypothesis tests - for these to be valid does the data in the sample have to be drawn from a bivariate normal distribution or does that only apply to the product moment correlation coefficient?

Cheers,
Just some guy

 

[EnumaElish]

As far as I understand, the data can be anything.

http://en.wikipedia.org/wiki/Spearman's_rank_correlation_coefficient (look under "Determining significance").

 

[Architeuthis Dux]

Hi there,

Spearman's rank correlation coefficient is a non-parametric measure of correlation between two variables, meaning it does not assume a specific distribution of the data. Therefore, the data does not have to be drawn from a bivariate normal distribution for the hypothesis tests to be valid. However, it is important to note that Spearman's rank correlation coefficient is best suited for monotonic relationships, not necessarily linear relationships. If the relationship between the variables is not monotonic, other measures of correlation may be more appropriate. I hope this helps clarify your question. Let me know if you have any further inquiries. Best,