RE: Help: Non-positive semi-definite message

From: "Piotrovskij, Vladimir [JanBe]" <VPIOTROV@janbe.jnj.com> Subject: RE: Help: Non-positive semi-definite message Date: Thu, 16 Aug 2001 15:58:41 +0200 Alan, >Now, one question, if we drop an eta from a parameter with covariates, (the addition of >covariates significantly decreases the eta value, from 0.4** down to 0.04**), does this >mean we push the interindividual variability be basically from the covariates? I think you are talking about OMEGAs, not ETAs (the latter are random variables, while OMEGAs are the corresponding variances). Your thinking about the role of covariates in the reduction of OMEGAs is correct. If no covariates are included in the model the interindividual variability is a pure random effect. Adding covariates reduces the random component of the interindividual variability. One can imagine the situation when all (or almost all) interindividual variability is explained by fixed effects of covariates, and there is no need in the random effect anymore. Best regards, Vladimir