# Thread: Factorial experiment in CRD with unequal replications

1. ## Factorial experiment in CRD with unequal replications

Hi guys, i am a graduate student in statistics.

i want to test for the main and interaction effect of 3 treatments on a group of patients. let say i have treatments A, B and C. each with 2 levels, presence and absence. So i decided to perform a 2^3 factorial Experiment in a Completely Randomized Design(CRD).
Now as the data arises from medical fields and test is not performed by me, i don't have equal no. of replications for all the 8 treatment combinations. And I cant find anywhere in my textbooks on how to deal with a Factorial analysis in CRD with unequal replications, and thus i can't find the sum of squares due to the treatment combinations and perform the tests. And also i want to know if i can apply standard Yates' algorithm for 2^3 factorial analysis.
Can you guys help me please? any help will be appreciated. Also if you can provide answers using R or MINITAB that is okay for me. But other softwares like SAS, i haven't used them yet.

2. ## Re: Factorial experiment in CRD with unequal replications

This is an unbalanced design. In Minitab, use GLM (General Linear Model).

3. ## Re: Factorial experiment in CRD with unequal replications

can you please elaborate the ANOVA table? i mean i can't understand how in Minitab they are obtaining the sum of squares due to the factorial effects. And so if you can explain them theoretically that will be great.

4. ## Re: Factorial experiment in CRD with unequal replications

In Minitab: Stat > ANOVA > General Linear Model > Fit General Linear Model > Help > See Also > Methods and Formulas > Adjusted Sum of Squares

 Tweet