Using vignette answers in a multinominal logistic regression

[Jeppe Poulsen]

Hi

I've made a vignette study, where there is 8 versions of the vignette, but respondents have only got 1 of them, where they have answered on a nominal scale.
There are 3 diffirent (gender, parents education and student grade) factors with 2 variations, which makes it 8 versions in total.

Is it possible to use stata's mlogit command to see which factor influence the answers the most? So I fx. can see if the gender is female instead of male, the likelihood of the respondents answering A is more likely than B (or C or D).

Thx in advance

Jeppe

[Phil Bromiley]

You'll increase your chances of a helpful answer if you follow the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

Multinomial logit sounds right, but you're probably going to want a multi-level model to control for the version of the vignette they answer.