Dummy coding and order of attribute levels

This forum is for posts that specifically focus on Ngene.

Moderators: Andrew Collins, Michiel Bliemer, johnr

Dummy coding and order of attribute levels

Postby mabbott » Thu Apr 13, 2023 12:37 am

I am trying to use dummy coding for the attributes in my DCE, and I'm wondering whether the order I list the levels in matters. For example, is:

b1.dummy[0.1|0.2|0.3] * Diagnosis[35,45,55,30]

The same as:

b1.dummy[0.1|0.2|0.3] * Diagnosis[30,35,45,55]

My DCE has three alternatives: A, B and Status Quo. The levels vary for A and B, but are always the same for the Status Quo alternative. I want to ensure that the correct level is applied to the status quo alternative.

Thanks,

Michael
mabbott
 
Posts: 6
Joined: Sat Nov 19, 2022 2:12 am

Re: Dummy coding and order of attribute levels

Postby Michiel Bliemer » Thu Apr 13, 2023 11:13 am

Yes the order in which you list the levels is important. But it does not depend on the level for the status quo alternative.

The LAST level in the list is always the reference level in Ngene, so when you write:

b1.dummy[0.1|0.2|0.3] * Diagnosis[35,45,55,30]

it means that:
level 30 has a utility of 0 (base)
level 35 has a utility of 0.1
level 45 has a utility of 0.2
level 55 has a utility of 0.3

Any of these levels can be fixed in the status quo, you can do this via:

;require:
statusquo.Diagnosis = 30 ? or statusquo.Diagnosis = 45

To be able to use require constraints, you need to use the modified Federov algorithm,

;alg = mfederov

Michiel
Michiel Bliemer
 
Posts: 1705
Joined: Tue Mar 31, 2009 4:13 pm


Return to Choice experiments - Ngene

Who is online

Users browsing this forum: No registered users and 12 guests

cron