Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Ratio of Dual Response None


This may be a simple question or may be more difficult than it seems at first.

I am wondering if it is possible to get the ratio of the dual response none?
I'm looking to see if I can use it as a secondary dichotomous none.

That is if the respondent were to choose either or of the nones the would still be indicating none but out of two types of none.

for example in the transport context, with Car, Bus, Train being ASCs with underlying attributes and levels, would it be possible to have none 1 - Bike and none 2 - Walk (to which no cost would be assigned in this instance) and then get the ratio of those who would walk and those who would ride?

Is this possible?

Many thanks.
asked Mar 6, 2017 by Jasha Bowe Bronze (1,745 points)
Perhaps my intentions are misguided or not. I want to really assess people's willingness to walk or ride (which I am, for a particular purpose not ascribing a monetary value, or other attribute / level as I am making the assumption that walking or riding is under the complete control of the respondent - this is an assumption which I am specifically making, and one could contend there are attributes and levels but for my study's purpose I choose them simply to binary). I want to keep them out of the choice task for brevity and width on screen etc. perhaps I should just use this question as a screener and terminate them prior to the choice task? but this feels like I might be missing some valuable data points?

1 Answer

+1 vote
Best answer

I think using just one None parameter and then computing e(U1-U2), where U1 is the ASC for bike and U1 is the ASC for ride will do the job for you (it computes the ratio of choice likelihood between bike and ride).  Estimating multiple alternative-specific Nones I think creates an identification problem.
answered Mar 6, 2017 by Bryan Orme Platinum Sawtooth Software, Inc. (184,140 points)
selected Mar 6, 2017 by Jasha Bowe
Oops, I meant to say e^(U1-U2)!

My guess is you weren't confused by my typo.