Need help with back-calculation of a correlation coefficient for a meta-analysis
I have never posted on this board before, so please bear with me if I am noob-ish.
I am doing a meta-analysis on plant physiology studies, using the correlation coefficients from studies that regressed plant traits against altitude. However, some give the response variable as SLA (specific leaf area, units m^2 g^-1) and others give it as LMA (leaf mass per area, units g m^-2).
SLA = 1/LMA
I want to convert to LMA, but many published studies do not include all the raw data. Usually they include some combination of the r^2, r, and/or b1 (regression slope) values.
So my question is, if you have the correlation coefficient r between two variables x and y, can you use that information alone to calculate r for x and 1/y? If not, what other information do you need?
Thanks so much for your help!
-Quentin @ Univ of Tennessee