I have three factors each at different levels. 1st factor has 2 levels (qualitative), 2nd factor has 4 levels (qualitative) and 3rd factor has 5 levels (quantitative). To further explain this experiment, the first two factors cannot be controlled only one factor(3rd) can be controlled.
Which type of design of experiments will best describe this type of experiment.
Difficult to answer, unless question is appropriately modified. Your question is contrasting, once you say one factor is uncontrolled, in description you say two factors are uncontrolled.
Cleopatra, try downloading the Quantum XL trial software package and use the QXL DOE/Create Design/ N-level factorial - this will give you the upper bound of full-factorial design. If you use the ACTS free-ware or rdExpert(lite) software for a HTT combinatorial solution (all 2-ways but hopefully the 'as orthogonal as possible' in rdExpert) then you have the least runs. D-optimal can reduce from full-factorial between the two extremes. Ultimately, before you run tests, populate the output response with random numbers & check orthogonality with your necessary interactions by examining the VIF or TOL values. Hope this helps, Keith
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling