I did a two-level full factorial design in order to find the significance of three factors (A, B, C) and their interactions on a response (R). Due to interaction between minimum level of A and maximum level of C at two runs, a situation turned up that R is not detectable. How can I treat with such a case? Shall I put these two runs away and analyze the remained ones? If so, what should I do with this non-orthogonal condition?
I wouldn't worry about it that much. Make sure your analysis conforms to the assumptions behind the type of model you make.
Was this expected?
What is your response? Is it categorical? continuous? count?
Try by simple T test,it is too lengthily to compare all possibilities,but confirms whether interaction is really significant or not.
Hello dear Hamed;
Which software you are using? DX7 (state-ease) software? Please send me your design and its relevant response maybe I can do something.
Dear Hamed,
Following my last comments, some times you can ignore some of your responses ( it means ignoring test and its responses; but keep in mind that in this case you will ignore its role in model for example ignore factor A; AB, C or BC not all of them!!) I think you should recognize your phenomena (I mean independent parameters and dependent variable; Is it any logical issue to accept this interaction AC?? (One non significant factor with significant factor)? For removing of bad data; you can check the predicted amount of model versus real data and check the errors and ignore bad data considering exceed levels. Also sometimes you can consider transform in your design. Which software you are using? I advise you to use Design-Expert® Software. However you can also use Minitab software
- Badges
- Science topic