choice-metrics.com

Hi,

I have a question about including overlap in attributes. I have previously posted a question regarding this design (viewtopic.php?f=2&t=695).
Using the following code below, we tested a few surveys via cognitive interviews and realize that a handful of respondents are focusing only on the overall survival (OS) attribute, and making choices based on OS only. Therefore, we would like to include some degree of overlap for OS i.e. having some choice tasks (but not all) that have the same level of OS for both alternatives. However, I'm not sure how to go about doing this. Because if I use
;require:
alt1.OS=alt2.OS;
that ends up with OS being the same for both alternatives across ALL choice tasks.

How can I set it up such that I have some choice tasks with overlap of OS, and some that does not - to try to address the issue of attribute dominance on OS?

Design
;alts = alt1*, alt2*
;rows = 52
;block = 4
;eff = (mnl,d)
;alg = mfederov(candidates = 5000)
;reject:
alt1.OS < alt1.TP,
alt2.OS < alt2.TP
;model:
U(alt1) = b1.dummy[0.002|0.001] * OS[22,16,12] ? overall survival (months)
+ b2.dummy[0.002|0.001] * TP[20,14,10] ? time to progression (months)
+ b3.dummy[0.002|0.001] * PAIN[100,50,25] ? decrease in pain (%)
+ b4.dummy[-0.002|-0.001] * FATI[2,1,0] ? fatigue [2] increase from mild to severe, [1] increase from mild to moderate, [3] remains at mild
+ b5.dummy[-0.003|-0.002|-0.001] * NAU[3,2,1,0] ? nausea [3] severe, [2] moderate, [3]mild, [0]none
+ b6.dummy[0|0] * ADMIN[2,1,0] ? administration, [2] oral, [1] Subcutaneous, [0] IV
/
U(alt2) = b1 * OS
+ b2 * TP
+ b3 * PAIN
+ b4 * FATI
+ b5 * NAU
+ b6 * ADMIN
$

Thanks
Xinyi

See Section 8.10 of the Ngene manual. Creating overlap across attributes in an advanced type of design called a partial profile design.
This involves creating allowable choice tasks in an Excel spreadsheet (the so-called candidate set) and then use:

;alg = mfederov(candidates = candidateset.xlsx)

Michiel