For my master thesis I made a Discrete Choice Experminent in JMP. The following design roled out of JMP (with 4 attributes, 3 levels each, and 16 choice sets)
When I checked the validity of the design I made a correlation matrix of dummy variables. The attributes Gar2/Gar3 (warranty) and EEO 450,525 and 600 (efficiency of energy conversion) correlates with each other. Is this correlation to high to be a orthogonal design or does it not reach a specific threshold/ boundary? In other words, do I need to make a different design?
