[R-meta] Question about obtaining correlation coefficient

James Pustejovsky jepu@to @end|ng |rom gm@||@com
Thu May 23 18:20:52 CEST 2019


If you only have the M and SD of each variable, then no it is not possible
to calculate their correlation.

If you have the above, plus the M and SD of the difference between the two
variables, i.e., mean and SD of (B - A), then you can back out the
covariance and correlation. Or if you have M and SD of the sum (A + B), but
it seems less likely that the sum would be reported.


On Thu, May 23, 2019 at 11:17 AM Cath Kids <cathkidslover using gmail.com> wrote:

> Hi everyone,
> I am currently conducting a meta-analysis on the association of two
> variables (bias A & bias B) in high anxiety individuals. Theoretically they
> should be related but needs to be verified quantitatively. I used mostly
> pearson correlation coefficients. But in some studies, while the two
> variables were measured, only their means and SDs were reported.
> My question is, is it possible to calculate the correlation between bias A
> and bias B with just the provided means and SDs without covariance
> (assuming both variables are normally distributed)?
> I understand that it is possible to convert *d *to *r ,* but does it apply
> to this situation if the two variables are measuring related but different
> concepts?
> Thank you very much!
> Regards,
> Joanne
>         [[alternative HTML version deleted]]
> _______________________________________________
> R-sig-meta-analysis mailing list
> R-sig-meta-analysis using r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis

	[[alternative HTML version deleted]]

More information about the R-sig-meta-analysis mailing list