[R] Significance of correlation

Rolf Turner rolf.turner at xtra.co.nz
Tue Jul 31 12:00:39 CEST 2012


On 31/07/12 21:13, Chintanu wrote:
> Hi,
>
> I understand that to test the significance of correlation between two
> PAIRED variables, the function, cor.test () can be used. However, in my
> case, I have tested the correlation (i.e., Correlation Coefficient, r)
> between two independent (i.e, different) variables, and now I wish to test
> for the statistical significance of the correlation. Could you please
> suggest how I should do that.

In a word, don't.  If two random variables are independent then they
are necessarily (and obviously) uncorrelated.

Note that "independent" is not synonymous with "different".

     cheers,

         Rolf Turner



More information about the R-help mailing list