Hey everyone,
I am new in this forum and it is my first time working with Stata and developing a Choice Experiment.
After defining my data:
- about 70 respondents, each of them answered 8 choice sets containing of 2 options and third an opt out alternative, resulting in 1680 observations
- 5 attributes, dummy coded, one of them is price
- unlabeled
I am wondering how I can proceed with running a logit model. I ran asclogit and mixlogit but I thought that I have to make tests in order to find out which model would be best or if I can say anything about IIA.
1. So as far as I understood, Likelihood ratio test would be an option to test wether asclogit or mixlogit fits better, is that true?
2. Futhermore, do I have to run a Hausman-Test and if so, am I testing the data itself or do I compare again between different models?
I read the descriptions for both tests in Stata but I am afraid I do not understand what the tests are exactly for.
Thank you
I am new in this forum and it is my first time working with Stata and developing a Choice Experiment.
After defining my data:
- about 70 respondents, each of them answered 8 choice sets containing of 2 options and third an opt out alternative, resulting in 1680 observations
- 5 attributes, dummy coded, one of them is price
- unlabeled
I am wondering how I can proceed with running a logit model. I ran asclogit and mixlogit but I thought that I have to make tests in order to find out which model would be best or if I can say anything about IIA.
1. So as far as I understood, Likelihood ratio test would be an option to test wether asclogit or mixlogit fits better, is that true?
2. Futhermore, do I have to run a Hausman-Test and if so, am I testing the data itself or do I compare again between different models?
I read the descriptions for both tests in Stata but I am afraid I do not understand what the tests are exactly for.
Thank you
