I have a two-level model with autocorrelated errors at level-1. Is there any way to specify this within proc mixed? I know there is a REPEATED option but the data are not panel data (e.g. only one observation per person).
As an alternative---any way to specify a multilevel feasible generalized least square model?
Any assistance is appreciated.
... View more