Hi, I have a question about how to do case weight in mixed model using R. I read Pinhiero and Bates (2000) Mixed-Effects Models in S and S-Plus. The weight functions in there are not what I wanted. Can R do case weight, giving each observation my own weight? If so, could you give me some references and examples? I'm looking forward to hearing from you! Tianyue
On Tue, 2 Sep 2003, tzhou1 wrote:> Hi, > > I have a question about how to do case weight in mixed model using R. I read > Pinhiero and Bates (2000) Mixed-Effects Models in S and S-Plus. The weight > functions in there are not what I wanted. Can R do case weight, giving each > observation my own weight? If so, could you give me some references and > examples? I'm looking forward to hearing from you! >It depends on what you mean by weights. The standard thing to mean in regression is weights that are inversely proportional to variances, but lme() does this, so I assume you mean something else. What? -thomas Thomas Lumley Assoc. Professor, Biostatistics tlumley at u.washington.edu University of Washington, Seattle
>It depends on what you mean by weights. > >The standard thing to mean in regression is weights that are inversely >proportional to variances, but lme() does this, so I assume you mean >something else. What?I want to use weights based on the residuals (not the usual residuals, refer to Markatou, M. (1996), ??Robust statistical inference: weighted likelihoods or usual M-estimation??? Communications in Statistics, Theory and Methods, 25, 2597-2613.). Basically, I want to calculate my own weights first, then use them in the mixed model in R. Can R do this? Thank you very much for your reply! Tianyue