the t-values for regression coefficients

[liamplayt]

How can I calculate the t-values of the regression coeffs if I have the sample size and the correlation matrix only?
Is this possible?

Thanks in advance.

 

[spunky]

if i also had the variance (or the standard deviation) of each independent variable i would be able to do it... but without those.... :S

 

[liamplayt]

that's what I thought as well...
what about this => t = r*sqrt(n-2)/sqrt(1-r^2)
r takes the partial correlation coeffs