[R-sig-eco] (no subject)

["Nils Gülzow"]

Dear List,

I have a problem with the ezANOVA package using for rm ANOVA or a problem with my statistic design of the experiment. (or  both ;-))

My file looks like this (is also attached):

sample	days	    treatments	         repl.	dispersal  gbiov/rich/evenness
1-36    7,14,21     synchron,asynchron   1-4    low,medium dependent variable
        28,35,42    controll                    high


I would like to compare within SSw (treatment,dispersal) over time (days= between, SSb)for the listed variables gbiov, rich, evenness (dv), of course a single analysis for each variable. I used the repl. column as the SS identifier for the replicates, so I have for each combination of dispersal x treatment 4 replicates.  

My problem is that for one combination (asynchron x low) I have only three replicates and by checking with ezDesign, I add the missing replicates by reseting to zero for each time point, but still I get error message, except when I add days to the within-SSw and omit between-SSb. But that makes no sence to me.

analysis<-ezANOVA(data=..., dv=.(gbiov), wid=.(repl), within=.(treat,disp),between=.(days))

Perhaps, the SS identifier makes no sence?!



Thanks in advance for solving the problem 

Best,
Nils 
--