Hello Colleagues,
By habit, I always check the variance of the residuals of my ordinary
regression models using: sigma(model)^2, which is also printed in the
output.
I know that the variance of the residuals in meta-regression is not
estimated but rather taken as being known and fixed by virtue of user's
supplying the 'vi' or 'V' to functions such as rma.uni() and
rma.mv().
So, I was wondering what is the interpretation of sigma(rma.uni_model)^2
and sigma(rma.mv_model)^2 and how they connect to the user-supplied 'vi'.
Thank you very much for your time.
Best,
Yuhang