I am sorry to ask a trivial question but I am not a statistician.
When I need to compare more than two groups in a unbalanced design with
SAS system I use PROC GLM (like the example in data5.csv from Cody R.
Applied statistics and SAS programming language p.223). R glm gives
different results.
In R,
-- lm fits ordinary least squares linear models; this is likely to
be what you want if you were using PROC GLM.
-- glm fits generalized linear models (e.g. logit, Poisson, Gaussian,
etc.).
-- lmer and lme fit mixed models, similar to PROC MIXED
Cheers,
Hank
On Jan 5, 2007, at 10:05 AM,