[R] linear model contrast questions

Jörg Schaber schaber at molgen.mpg.de
Mon Sep 24 20:13:44 CEST 2007


when I fit a linear model as

 lm(o ~ y + s - 1, contrasts=list(s=("contr.sum")),na.action=na.exclude)

should then the effects of factor s sum up to zero?

I my case they don't. Where is the misunderstanding or the mistake?



More information about the R-help mailing list