choice-metrics.com

Hi all,

I'm a beginner. I've 2 questions here; 1) is it normal for Ngene to take more than a day to generate the choice set? currently the iteration is keep running and I'm not sure when it will stop. 2) I didn't include the socio-economic variables in the utility function, will this be alright?

Here is the syntax.

Design
;alts = A*, B*, Default
;rows = 12
;block = 4
;eff = (mnl,d)

;model:
U(A) = b1[0]
+b2 [0] * waste[0,20,40]
+b3 [0] * time[0,10,20]
+b4.dummy[0I0] * bin[1,2,3]
+b5[0] * pay[0,5,100,250]
+b6[0] * benefit[0,4]/

U(B) = b2 * waste
+b3 * time
+b4.dummy*bin
+b5* pay
+b6* benefit

$

When generating orthogonal or random designs, Ngene will automatically stop once it has found a design. However, when searching for efficient designs, there are billions and billions of possible combinations, so the search process WILL NOT STOP UNLESS THE USER PRESSES THE STOP BUTTON. You can stop of Ngene is not able to find a better design after a few minutes or so. You will see that Ngene at some point will no longer produce designs that have a substantially lower D-error, the D-errors will more or less be the same and therefore you can stop the search algorithm and use the best design it has found so far. There is not much benefit in waiting for a very long time to find a design with a marginally lower D-error.

To answer your second question, yes that is correct, you typically do not include socio-demographic variables in your design phase since you do not yet know who your respondents will be. You will add these into the utility function after you have completed your data collection and you have the data about your respondents.

Michiel

Thank you for the explanation Michiel. Really appreciate it.