cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Choose Language Hide Translation Bar
NanoK
Level I

Help with Anova, Effects, and Multiple comparisons

Hello, 

 

I am trying to run an Anova on a data set I have from our lab. Briefly, we exposed some particles at 3 different sizes to some cells at 2 different doses and measured gene expression. 

 

The issue I am running into is that when I go to perform my Anova it keeps giving me an error for my Dose and won't give me a P-value for it. As a result I cannot run a Tukeys-HSD post-hoc for my multiple comparisons and look at particle*dose interactions. 

 

I have my Gene Expression as Numeric/Continuous, Particle Size as Character/Nominal, and my Dose as Numeric/Continuous. For Model Dialog I have my Gene Expression as my Y variable and Particle Size, Dose, and Particle Size*Dose as my model effects. 

 

Based on these parameters I get the following output (see screenshots). I have tried several variations of this setup and even attempted to solve the problem using R and myself/others get the same results.

 

Any help is greatly appreciated.

Thanks

1 REPLY 1

Re: Help with Anova, Effects, and Multiple comparisons

It appears as though the model cannot estimate the interaction term because your control was not run at multiple doses like the other levels of particle size. I see three possible courses of action:

1) Remove the interaction term from the model

2) Remove the control runs from the data by excluding them (and maybe hiding too, but maybe not hiding depending on what you want to see)

3) Add different dosage levels of your control to the data table.

 

The course of action you take will depend on the information you wish to obtain on the control particle size compared to the other particle sizes.

Dan Obermiller