choice-metrics.com

Dear all,

First of all, sorry if I am asking some silly questions but I am sure this forum (in particular the moderators) will give me some fruitful advice as usual. I want produce an unlabeled DCE design which has three alternative where the last one is a no-choice alternative. There are two (4 and 6 levels each) attributes which are specified as linear in the utility function. Since I have only two attributes that are linear, I want to use the orthogonal design instead of the efficient design so the attribute levels are balanced and they are fully independent from each other. I generate a design with 12 rows (see below including the design). The number of levels in each column are not equal - I know this is because it is a fractional factorial orthogonal design (FFOD), but the issue I am worrying about is that when these design combinations are presented to respondents, they (respondents) do not see the same number of attribute levels in each choice situation. So, how can we ensure that the number of levels are presented the same number of times in each choice situation when using an FFOD? Or is that not a problem at all? I raised this question because I thought one aim of using FFOD is to obtain the same number of levels in each columns. Thank you indeed for your time!

Design
; alts = alt1, alt2,alt3
; rows = 12
; fact
; Model :
U(alt1) = amo*amount[1,2,3,4]+cost*price[1,2,3,4,5,6]/
U(alt2) = amo*amount+cost*price$


Choice situation alt1.amount alt1.price alt2.amount alt2.price
1 2 5 1 1
2 3 3 2 3
3 3 5 4 6
4 4 4 1 5
5 4 1 2 3
6 3 3 4 4
7 3 5 2 1
8 3 6 1 6
9 4 5 3 4
10 1 4 4 4
11 4 1 4 2
12 4 3 2 1

You need to put in the ;orth command to generate an orthogonal design and not the ;fact command, which generates a random design. Please see the manual for the options on ;orth

Dear Michiel,

Thanks for the correction - I will run the design using the ;orth property and see what happens. Thank you indeed!