Dear subscribers
I performed a quite simple pollination experiment:
12 plants
12 pollination treatments per plant,
plant 1 had treatments a, b, c,
d... etc plant 2 had treatments a, b, c, d...
The idea was to run a simple anova using plants as blocks to see the effects of treatments on fruit production, but I lost some a few treatments in some plants, thus I have to deal with missing values
First I tried the obvious, arc transformed fruits
fruset.aov<-aov(arcfruit~treatment+Error(Plant),
na.action=na.omit, fruset)
but I get the wrong DF because of the missing values, and I cannot perform the multiple comparisons test. Is there a better way to deal with this unbalanced design?
I did a little bit more research and I decided to run a GLM for mixed effects using a binomial distribution
y<-cbind(fruits,nofruits)
model<-glmer(y ~ treatment+ (1|Plant), binomial,data=fruset)
but them I cannot get the anova table
Error in anova(model, test = "F") :
single argument anova for GLMMs not yet implemented
Some people suggested using Anova function on the package car, but I don't see how can I get it to work with a mixed effects model like this.
Any suggestions are appreciated. Are there other straight forward ways to analyze such data, given the missing values, and multiple comparisons follow up?
Thank you
Humberto
==========================================================
'Discipline - Success doesn't just happen. You have to be intentional about it, and that takes discipline.' - John Maxwell
[[alternative HTML version deleted]]