Dear all;
I'm looking for some help in translating the following SAS code to R. The
code represents a
factorial design plus 1 control plot (2 x 2 + 1). The data is the following
BLOCK FA FB FC Y
1 0 0 0 15.33
1 1 0 0 14.40
1 1 1 0 15.49
1 1 0 1 16.87
1 1 1 1 18.84
2 0 0 0 15.97
2 1 0 0 16.18
2 1 1 0 14.52
2 1 0 1 18.04
2 1 1 1 19.81
3 0 0 0 15.60
3 1 0 0 14.79
3 1 1 0 14.30
3 1 0 1 17.18
3 1 1 1 18.37
4 0 0 0 15.22
4 1 0 0 16.24
4 1 1 0 15.97
4 1 0 1 16.51
4 1 1 1 19.05
The SAS code is:
proc glm data=test;
class BLOCK FA FB FC;
model Y = BLOCK FA FB(FA) FC(FA) FB*FC(FA)/alpha=0.05;
random BLOCK/test;
lsmeans FB*FC(FA)/pdiff stderr;
quit; run;
I've tried:
library(nlme)
options(contrasts=c("contr.SAS","contr.SAS"))
data.gr<-groupedData(Y~1|BLocK,data=data)
data.gr$BLocK<-as.factor(data.gr$BLocK);data.gr$FA<-as.factor(data.gr$FA)
data.gr$FB<-as.factor(data.gr$FB);data.gr$FC<-as.factor(data.gr$FC)
mod1<-lme(Y~FA+(FB*FC)/FA,random=~1|BLOCK,data=data.gr)
but I can't get the model works and I can't figure out how can I do it.
I'll appreciate any ideas
Best regards
P. Porras