Page 1 of 1

Attributes levels estimation

Posted: 20 Nov 2022, 05:08
by halilka
I conducted a discrete choice survey in which each respondent had to choose 1 set out of 2 choice sets (Car, Bus), which are characterised by 4 attributes (let's say price, time, distance, day) of either 2 or 3 levels.

I tried to calculate the utilities/coefficients with the estimation of multinomial logit model using the apollo package in R. It's working well to calculate the utilities for the attributes.

But I want to calculate the coefficients not for the attributes, but for the levels. I want to know the utility, e.g. for the time "15 minutes". How can I calculate this? Is it possible at all to calculate it with the apollo package in R?

I am grateful for every piece of advice!

Re: Attributes levels estimation

Posted: 25 Nov 2022, 13:14
by stephanehess
Hi

this is straightforward, please see the example MNL_SP and the treatment of the attribute service_quality in that example

Stephane

Re: Attributes levels estimation

Posted: 26 Nov 2022, 15:58
by halilka
Thank you so much for your kind reply.