送交者: xj 于 2005-8-20, 11:33:00:
回答: Sure. Is it worthwhile at the same time to do a linear 由 cornbug 于 2005-8-20, 01:22:05:
but do a repeated measures model for each variable using mixed model is preferred.
I am worrying about SAS GLM repeated method. I heard rumors that it is wrong. I guess maybe because GLM model all variables as a mutivariate normal distribution, which is not necessary true.
But most models are pretty robust about normal distribution stuff if you have large sample size.