The first model includes C and interactions with C, which are combined
with the residuals in the other two models. The first model has 3, 3,
3, 3, and 32 degrees of freedom for C, A:C, B:C, A:B:C, and residuals,
respectively. The other two models combine these two into a single
residual term with 44 = 3 + 3 + 3 + 3 + 32 degrees of freedom.
hope this helps. spencer graves
Cec?lia Shiraiwa wrote:
> Hello all,
> I?m trying to study a factorial design, but I can?t understand why did
Df, Sum Sq and Mean Sq of residuals alter when I Split the interaction? I think
that Split the interaction must not alter the residuals. Am I doing something
wrong?
> Could anyone help me?
> My data and functions I tried are:
>
>Y<-c(196,213,183,
> 192,253,199,
> 251,331,276,
> 128,220,196,
> 272,275,227,
> 204,305,185,
> 135,322,276,
> 262,284,250,
> 272,275,227,
> 204,305,185,
> 135,322,276,
> 262,284,250,
> 296,369,344,
> 325,396,403,
> 370,323,319,
> 341,418,318)
>A<-factor(rep(c(0,1),c(24,24)))
>B<-factor(rep(c(0,1,0,1),c(12,12,12,12)))
>C<-factor(rep(c(rep(0:3,each=3)),times=4))
>av <- aov(Y ~ A + B + A*B + C + A*C + B*C + A*B*C )
>summary(av)
>avAB <- aov(Y ~ A/B)
>summary (avAB, split=list("A:B"=list(A1=1, A2=2)))
>avBA <- aov(Y ~ B/A)
>summary (avBA, split=list("B:A"=list(B1=1, B2=2)))
>
>
>Thanks
>Cecilia
>
> [[alternative HTML version deleted]]
>
>______________________________________________
>R-help at stat.math.ethz.ch mailing list
>https://www.stat.math.ethz.ch/mailman/listinfo/r-help
>
>