84083_en-US

liliconcom

ANOVA fit model

hi

I am comparing yield of three varieties of potatoes, variety A has 10 samples, variety B has 5 samples and varietry C has 20 samples. So sample number (replications) are not even. I want to perform ANOVA using fit model. I have created three comuns Variety, sample#, and yield.

So, yield (continuous) is the response factor. In the constrct model, do I have to include only varietry (nminal) or both varietry (nominal) and sample (nominal).

Thanks

Re: ANOVA fit model

You do not need to explicitly include Sample in the model. It is implicitly used to estimate the (sample) error (residual).

 

You can use Analyze > Fit Y by X with Yield > Y and Variety > X and then select ANOVA from the platform menu (red triangle). You can also use Analyze > Fit Model with Yield > Y and Variety > Effects. The resulting Oneway and Fit Least Squares platforms share some common information but each provide unique information so try them both.