In R package "psychometrics" an estimate of SE of R squared of

sersq <- sqrt((4*rsq*(1-rsq)^2*(n-k-1)^2)/((n^2-1)*(n+3)))

with n sample size, and k number of parameters if sample size greater than 60 is found. Does anyone have a formula for smaller sample size or an exact formula?

I have been through sos - but only found the above? - the real underlying problem is whether comparing models aiming at explaining here a series of concentrations based solely on the R-squared of the individual model predictions vis a vis the measured values is a healthy method



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.