D-efficient design dominant alternatives?

This forum is for posts that specifically focus on Ngene.

Moderators: Andrew Collins, Michiel Bliemer, johnr

D-efficient design dominant alternatives?

Postby andreb84 » Thu Feb 27, 2020 1:46 am

Dear all,
I've generated the following efficient design. I don't know the priors but I know the sign so I put a small number as priors:
Code: Select all
Design
;alts=SQ, alt2, alt3
;rows=10
;eff=(mnl,d)
;model:
U(SQ) = b1[-0.0001] * TAX[109] + b2[0.0001] * PROD[60] + b3[0.0001] * PER[100] + b4[0.0001] * CHIL[100]/
U(alt2) = b1 * TAX1[25,50,109,200,300] + b2 * PROD1[30,60,90,120] + b3 * PERF1[50,100,150,200] + b4 * CHIL1[50,100,200,300] /
U(alt3) = b1 * TAX1 + b2 * PROD1 + b3 * PERF1 + b4 * CHIL1
$


However, In some choice task I have some clearly dominant alternative, for example in one case the two alternatives have a greater level of TAX (which has a negative prior) than the status quo and a lower level of all the other attributes (which have positive prior) than the status quo.
Is this possible? Is because the priors are very small number? What do you suggest?
Thank you for your attention
andreb84
 
Posts: 6
Joined: Tue Dec 03, 2019 10:25 pm

Re: D-efficient design dominant alternatives?

Postby Michiel Bliemer » Thu Feb 27, 2020 8:30 am

You need to tell Ngene which alternatives are generic/unlabelled such that it can check for dominance. You need to set the following:
;alts=SQ*, alt2*, alt3*

If the signs of your priors are set correctly, this will remove all dominant alternatives in your choice tasks.

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

Re: D-efficient design dominant alternatives?

Postby andreb84 » Thu Feb 27, 2020 9:03 pm

Thank you very much Michiel!
andreb84
 
Posts: 6
Joined: Tue Dec 03, 2019 10:25 pm


Return to Choice experiments - Ngene

Who is online

Users browsing this forum: Bing [Bot], Google [Bot] and 8 guests

cron