Page 1 of 1

reference alternative with attributes?

PostPosted: Wed Jun 08, 2011 6:46 am
by Sanchez
As mentioned in a previous post, our project will be examining the visual disamenities of structures which are at different distances from a touristic viewpoint (there is a fee to access this viewpoint). The number / size of the structures is also varied. Respondents are asked what price discounts they would be willing to accept to face the new view (with structures in the viewshed). As such, my three variables are distance (dist: 5,8,12,18miles), size (small and large) as well as price (+5%, 0%, -5%, - 10%, - 15% -25%).

I can’t quite figure out how to deal with our opt-out option from the manual. Our opt-out (status quo) option is actually a reference alternative. The reference alternative has a specified distance (30 miles), size (144) and price (0 - zero percentage change). As you can see, the levels associated with the reference alternative do not occur in the choice alternatives (except the price level).

The design we are thinking of using in our pilot study to generate priors is the following

Design
;alts = alt1,alt2,sq
;rows=12
;block=2
;eff=(mnl,d)
;model:
U(alt1) = b2.effects[0|0|0]*dist[8,12,18,5] + b3.effects[0]*size[144,64] +b4[0]*price[-5,5,10,15,25,0] /
U(alt2) = b2*dist + b3*size +b4*price
$

The problem is, that I read in the manual that Ngene assumes utility set to zero for this alternative and that there are no attribute levels attached to the alternative.

So – should I specify more information about the sq alternative? (see for example below)?

Design
;alts = alt1,alt2,sq
;rows=12
;block=2
;eff=(mnl,d)
;model:
U(alt1) = b2.effects[0|0|0]*dist[5,8,12,18] + b3.effects[0]*size[64,144] +b4[0]*price[0,-5,5,10,15,25] /
U(alt2) = b2*dist + b3*size +b4*price /
U(sq) = b2[0]*dist[30] + b3[0]*size[144] + b4[0]* price[0]

$

I forsee a problem however in using effects coding for b2 in the first two alternatives, but then not for U(sq)?

How do I handle this problem? Can I somehow use the first design specified above to obtain priors and then somehow infer a value of the U(sq) prior for this assuming that the utility for the reference alternative will be fixed (but not zero)? If so, how?

Moreover, when I run the code as specified in the first coding above, the design looks good, but I noticed that in all the choice questions, the size is always different for the two alternatives offered in the choice (i.e. The size is always either 64 in alt1 and 144 in alt2 or the other way around. The size is never 64 for both alt 1 and alt2).
When I run the same model, but taking out the effects coding then I do get choices where the size is either 64 or 144 for both alternatives.
Why is this?

Any assistance would be much appreciated.
Kind regards
Sanchez.