Dear Sawtooth Team,
I would like to use the estimates from an alternative-specific design CBC for constructing a simulation tool. I used the MBC simulator as a template and have done so successfully for other CBC studies.
My question:
What do I have to consider for alternative-specific CBC?
I assume that the only difference would be in the model tab.
There is one primary attribute (PA) which corresponds with the "brand"/choice options. For six of the PA levels their is additional secondary attributes, for the other five there is none.
Will the model have eleven lines (for each level) plus one for "none"?
Will these lines include only a single 1 at the corresponding level of the PA when no other levels apply, ie. act like an alternative-specific constant?
Is the none-line all zeros as usual, i.e. nothing applies, except for the none utility?
Another question concerns the utilities in themselves. How can I interpret the relative importances, specifically that of the PA?
Thanks for some pointers!
Alex
