Reply to "annamb5" <anabras@...>'s questions.


Q1.Supposing that factor interactions were not included in the
design.
When doing ANOVA table (with QT4 soft)does the error percent
influence
also combines the effects of those interactions?

Ans. Yes. The effects of interactions will also be combined with the
effects you find for factors. This would be very common unless you
can afford to run full-factorial tests.

Q2.In the previous situation, if the confirmation test result (to the
optimum solution)is within the confidence interval does it means that
there are no interaction factors in that situation?

Ans. If the mean of samples tested at the optimum condition
(confirmation tests) falls within the confidence interval, it does
not imply that there is no interaction. It simply means that you rae
within your predicted performance in the presence of interaction.

Consequently, it means that the interactions are small and/or they
nullify each other resulting in total effects that is small.


Q3.When applying the automatic test of the presence of interactions
how
can I know what factors are significant or not?

Ans. You have no way to know that. Only way to learn about the
relative influence of interaction is to sacrifice column to study
interaction. So, what does the SI-interaction (in Qualitek-4
software, www.nutek-us.com/wp-q4w.html ) information tell you or when
can you use it? You can use this information only for future repeat
tests in deciding which interactions to study should you have emplty
column to study interaction.

Hope this would be of help to you.

Good luck.

- RKR 9/9/08