I'm trying to analyse data from a factorial experiment with missing cells. I've read this reference http://users.monash.edu.au/~murray/BDAR/ which recomends to treat the experiment with an one-way anova, and make contrasts to evaluate the main effects and interactions.
following the book examples I did
contrasts(fdata$ALL) <- cbind(c(rep(1,46),rep(-1,45), rep(0,32)),c(rep(-1,46),rep(0,45), rep(1,32))) AnovaM(aov(resp ~ ALL, fdata), split = list(ALL = list("PIE" = 1:2, " PIE1 vs PIE2"=1, " PIE1 vs PIE3"=2)))
I'm having trouble generating contrasts in R, and knowing how to generate the correct contrasts to each interaction.
I have six factors, five with 3 levels, and one with two, does any one know a systematic way to treat this kind of problem?