[R] help with aov

Martin Olivier martin at lirmm.fr
Thu Mar 18 19:17:10 CET 2004


Hi all,

Suppose the following data and the simple model
y<-1:12+rnorm(12)
fac1<-c(rep("A",4),rep("B",4),rep("C",4))
fac2<-rep(c("D","C"),6)
dat<-data.frame(y,fac1,fac2)
tmp<-aov(y~fac1+fac2,dat)

the command tmp$coeff gives the fllowing results :

(Intercept)       fac1B       fac1C       fac2D
3.307888    2.898187    7.409010   -1.088588

But mean(y) gives 6.199327 and is different of Intercept..
So, I don't understand the parametrization with the aov() function.
I would like the estimations with the constraints
 fac1A+fac1B+fac1C =0 and fac2D+fac2C=0...

What is the solution?


Thanks,
Olivier.




More information about the R-help mailing list