D-efficiency of orthogonal design

This forum is for posts that specifically focus on Ngene.

Moderators: Andrew Collins, Michiel Bliemer, johnr

D-efficiency of orthogonal design

Postby wxy008 » Fri Aug 27, 2021 12:59 am

Dear Michiel,
I run an example of labeled orthogonal design in the Ngene manual (see below). I did not find the D-efficiency value. I could get the D optimality by click OOD box on the left side of Ngene result page. I think the D optimality is not D efficiency for labeled design since we can not use OOD for labeled design. Is there any way I can get D-efficiency?

Thanks for the help.

Wei
Code: Select all
Design
;alts = alt1, alt2
;rows = 4
;orth = sim
;model:
U(alt1) = b1 + b2 * A[0,1] + b3 * B[0,1] /
U(alt2) = b2 * A + b3 * B
$
wxy008
 
Posts: 33
Joined: Wed Mar 04, 2020 3:22 am

Re: D-efficiency of orthogonal design

Postby Michiel Bliemer » Fri Sep 03, 2021 11:40 am

D-efficiency is only available for unlabelled experiments due to the very specific way that these designs are constructed using the algorithm in OOD and the formulas in the work of Street, Burgess, and Louviere. For labelled experiments these formulas no longer work.

The utility functions in this example may actually indicate an unlabelled experiment because the attributes are identical across alt1 anf alt2, the attribute levels are identical, and the coefficients for the attributes are identical. Therefore, if you change ;orth = sim to orth = ood you will get a 100% D-efficient design for this example (ignoring the constant).

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

Re: D-efficiency of orthogonal design

Postby wxy008 » Tue Sep 07, 2021 10:35 pm

Thanks so much!
wxy008
 
Posts: 33
Joined: Wed Mar 04, 2020 3:22 am


Return to Choice experiments - Ngene

Who is online

Users browsing this forum: No registered users and 6 guests

cron