Thank you for your response Michael, but I'm still looking for feedback on the proposed method that I outlined in the original post.
Specifically,
1) Does it make sense to use smoothest to estimate the smoothness on the level1 residuals and then average the DLH and RESELS from all those estimates?
2) When I use smoothest on the level 1 residuals should I use the number of time points to determine the dof?
thanks to all
|