Subscribe Bookmark RSS Feed

constraints on 3 profile choice sets

Hi,

I am trying to set up a discrete choice experiment that compares 3 different drink types, and I would like each choice set to include three options - drink 1, drink 2, and drink 3 (whilst varying the attributes of each drink).

Could anyone advise me on how this might be done in JMP?

Thanks

Jon
8 REPLIES
louv

Staff

Joined:

Jun 23, 2011

Hello

Have you checked out the demo tutorial at the following link

http://www.jmp.com/software/jmp8/demos.shtml
Hi,

I looked at the two design of experiment videos, but did not see anything relating to restricting combinations of attributes when defining a choice set of more than 2.

Do you know who I might contact at JMP with such a question?

Regards

Jon
louv

Staff

Joined:

Jun 23, 2011

Have you considered using the Custom Design feature and leveraging disallowed combinations? If you go to the JMP help menu and utilize the search feature and type "disallowed combinations" it discusses this capability.

If however you are set on a choice design or conjoint analysis then check out Help>Books>JMP DOE Guide>Discrete Choice Designs. Chapter 8 discusses specifying prior means and prior variance.
I hope this helps.
herberthinterbe

Community Trekker

Joined:

Oct 11, 2014

Hi LouV,

"If however you are set on a choice design or conjoint analysis then check out Help>Books>JMP DOE Guide>Discrete Choice Designs. Chapter 8 discusses specifying prior means and prior variance.

I hope this helps."

Unfortunately the information about the specification of the prio means and prio variance are quite poor. After reading all the information provided by JMP I actually still don't know how to use them. Is there any good example available? (Optimal Design of the Choice Experiment - JMP Blog is not really helpful as the author does not tell why he uses the values for the prio mean and prior variance.)

Best regards,

Herbert

Hi Lou

Thanks for your suggestions. I do want to undertake a choice experiment, but think I may have found the answer, by creating a table without the factor that I want to control and then adding in a new column with the factor in sequence.

Regards

Jon
Dear Lou V,

can you please elaborate on the subject of excluding factor combinations specifying prior means and prior variance? There is nothing in the help or handbooks regarding that topic.
More specifically: What value should be typed in the matrix if one wishes to disallow a certain factor combinations?
Thank you in advance for any help.

Regards
Marcel
choice_econ

Community Trekker

Joined:

May 11, 2012

To impose a prohibition (aka: constraint, disallowed combination, etc.), for instance disallowing level 1 across attributes X1 and X2, in the in the Disallowed Combinations dialogue, use the following: X1 == 1 & X2 == 1 .

This same format can handle other (non-equal) prohibitions, as well as multiple attributes (for instance across three attributes as X1 == 1 & X2 == 1 & X3 == 1).

Separate prohibitions using a semi-colon.

Anyone?

Please surprise me by proving that this forum is actually useful. So far I never got an answer to any of my questions.