Subscribe Bookmark RSS Feed

DoE - Discrete choice - Dominated choice sets?

ac11ca

Community Trekker

Joined:

Mar 2, 2012

Hi,

I am trying to design a discrete choice study. I have 2 variables each with 3 ordered levels. I have followed the “Discrete ChoiceDesigns” tutorial and used data from a preliminary pilot experiment to specifythe prior mean and variance matrix to indicate the order of the variablelevels. Nonetheless, the DoE module produces choice sets that are completely dominated, that is, one choice has all the best features and the other choicehas all the worst features. This is clearly not useful for the analysis.

Any help would be appreciated understanding what is going on.


Cheers,

Adrian

1 REPLY
martindemel

Staff

Joined:

Jul 10, 2014

Hi Adrian,

even you have done some Pilot Experiment it may is helpful to take a look on the recommendations in this thread Prior mean and prior variance matrix in choice design to check if the values you've entered are good values.

Best,

Martin