Skip to content
Prev 14046 / 20628 Next
Original Raw

GLM: Difference between treatment groups for each colony (Tukey posthoc test)

paul debes 
Hi Sophie and List,

If you are interested in the treatment contrasts for each level of the  
Colony factor (with three levels), it may be better to view your  
experiment as a factorial and specify a simple linear model rather than  
the presently specified linear mixed model. A random Colony term with only  
three levels may also not provide the best estimate for the correlation of  
colony data between treatments (i.e., as intraclass correlation) that is  
taken into account when testing the general treatment term.

Would this work for you?:

fit_life = lm(last.scan ~ 1 + treatment*colony, data = data)

In case the treatment:colony term is significant you could conduct the  
additional pairwise tests of interest.

Best,
Paul



On Mon, 07 Dec 2015 13:02:36 +0200, Sophie Waegebaert
<sophie.waegebaert at gmail.com> wrote:

            

      
      
      
    

        

      
      
    

        

  
    

      
      
    

  


  


  


  
  

    

  

    
      
    
    
Thread (5 messages)

  

  

    
            
        Sophie Waegebaert
        
        GLM: Difference between treatment groups for each colony (Tukey posthoc test)
      
      Dec 7
    
    
      
            
        paul debes
        
        GLM: Difference between treatment groups for each colony (Tukey posthoc test)
      
      Dec 7
    
    
      
            
        Sophie Waegebaert
        
        GLM: Difference between treatment groups for each colony (Tukey posthoc test)
      
      Dec 7
    
    
      
            
        paul debes
        
        GLM: Difference between treatment groups for each colony (Tukey posthoc test)
      
      Dec 7
    
    
      
            
        Peter Claussen
        
        GLM: Difference between treatment groups for each colony (Tukey posthoc test)
      
      Dec 7