Page 1 of 1

General choice modelling question

Posted: 19 Dec 2024, 14:56
by paul
Hi,

I would like to ask a question that does not relate to Apollo itself but to choice models and their evaluation in general. I hope you can help me anyway.

We want to analyse scenarios (let's call them x,y,z). Suppose in our choice experiment there are 3 attributes A,B,C, each with three levels 1,2,3. Scenario x has the levels lets say 1,1,2 for the attributes A,B,C, scenario y has the levels 1,3,2 etc. Is it methodologically possible to simply sum up the coefficients of the corresponding levels calculated with our MMNL model in order to calculate a total value for each scenario? Do you know of any literature that has already done this?

Thank you very much!
Paul

Re: General choice modelling question

Posted: 19 Dec 2024, 20:25
by stephanehess
Hi

this wouldn't really give you a lot of information as utilty is on an arbitrary scale

Stephane