Page 1 of 1

Positive alternative specific constant for the opt-out

Posted: 25 Oct 2022, 09:56
by DavidKL
Hi Apollo team!

I had a pilot study with 54 respondents and I got an interesting results. I have 8 decision situations with 3 alternatives and an opt-out option.

I got the following results for the choices:

Overview of choices for MNL model component :
alt1 alt2 alt3 altnc
Times available 432.00 432 432.00 432.00
Times chosen 149.00 108 134.00 41.00
Percentage chosen overall 34.49 25 31.02 9.49
Percentage chosen when available 34.49 25 31.02 9.49

so the no choice alternative was selected the fewest times.

But at the same time I get negative and significant asc for the opt-out alternative, which (if I am right) means that opt-out is more preferred than choice an alternative.

What can this ambiguity refer to?

Thank you for the answer!

David

Re: Positive alternative specific constant for the opt-out

Posted: 25 Nov 2022, 13:33
by stephanehess
David

a positive ASC simply means that "all else being equal", this alternative is preferred, i.e. if all attributes were the same, etc. You clearly have other terms in the utility function that capture the other element of the preferences

Stephane