covariate model building: forward inclusion and backward elimination

From: nonmem@optonline.net Subject: [NMusers] covariate model building: forward inclusion and backward elimination Date: Mon, 30 Oct 2006 10:22:30 -0500 Hello NONMEM users, This is question about forward inclusion and backward elimination. Obviously, backward elimination should be done using NONMEM. Some scientist apply NONMEM to do forward inclusion. Some other scientists use SAS or S-plus to run regression analysis on Bayesian estimates while including covariates in the model. Is that appropriate to use SAS or some other software instead of NONMEM during the forward inclusion step? Thank you, Pavel _______________________________________________________