RE: OMEGA HAS A NONZERO BLOCK

From:"Serge Guzy" Subject:RE: [NMusers] OMEGA HAS A NONZERO BLOCK Date:Tue, 8 Oct 2002 09:07:46 -0700 I think Leonid summarizes very weel the potential scenarios (I talked about also) we could encounter. I personally simulated the two scenarios he discussed which include high correlation, small standard error and shallow OF profile with correlation characterized by high standard error. I still suggest to use the bootstrap rather than the Hessian approach. It is may be time consuming but can prevent wrongly standard error estimation. Serge Guzy Head of Pharmacometrics Xoma