[R-sig-ME] Random factor and MANOVA

Aitor Larrañaga aitor.larranagaa at ehu.es
Wed Jan 19 23:13:25 CET 2011

Dear all,

I have an experimental design with two fixed factors and a blocking factor (random) and I have data on several variables for which I want to perform a MANOVA. 

After creating a matrix with the variables I want to be included in the MANOVA with 

>Y<-cbind(variableA, variableB... variableN), 

I ask for the linear model 


What creates an output that seems to be correct for a MANOVA. 

          			Df   	Wilks 		approx F 	num Df 	den Df    Pr(>F)    
factor1         		1 	0.61909   	7.2808     	6     		71 		4.393e-06 ***
factor2         		1 	0.92156   	1.0072      	6     		71    	0.4277    
factor1:factor2     	1 	0.90012   	1.3131      	6     		71    	0.2627    
Residuals 76                                             
Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

Nevertheless. I've been told that random factors cannot be included in a Manova models. Is this true? And if it is, what is the reason and what does R do with the model I've written then?

Many thanks,


More information about the R-sig-mixed-models mailing list