Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: how to derive standard error of correlation coefficient


From   Nick Cox <n.j.cox@stata.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: how to derive standard error of correlation coefficient
Date   Tue, 04 Aug 2009 08:52:59 -0500

If that is really all you know, I doubt that you can do it. To a good first approximation the se of r depends mainly on the sample size, so long as correlations are near zero. The original standard deviations are immaterial, given that the correlation is necessarily scale-free. But even given a p-value, you need sample size as well.

Also watch out: if correlations are interestingly non-zero, then the usual kind of rule that uncertainty is captured by intervals of the form estimate +/- multiplier * se breaks down, as the bounds +1 or -1 impart asymmetry to the problem. It's better to do calculations on a transformed scale. For more, see

SJ-8-3 pr0041 . Speaking Stata: Corr. with confidence, Fisher's z revisited (help corrci, corrcii if installed) . . . . . . . . . . . . N. J. Cox
        Q3/08   SJ 8(3):413--439
        reviews Fisher's z transformation and its inverse, the
        hyperbolic tangent, and reviews their use in inference
        with correlations

Nick

Miranda Kim wrote:

How can I derive the standard error of the correlation coefficient when I have only a correlation coefficient, p-value, and the standard deviations of both variables?

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index