How can i produce Odds ratio using mlogit command?

[Pawan Acharya]

Hello there,
svy: mlogit depvar i.indepvar , rrr
produces the estimates as the relative risk ratio. I am interested to get Odds ratio. Is there any way i can produce odds ratio? Suggestions will be highly appreciated. Thank you.

[William Lisowski]

How do you think the odds ratio would be defined?

For logisitic regression, where there are two outcomes, a and not-a, the odds ratio is the ratio of the odds of a to the odds of not-a.

For multinomial logisitic regression, there are more than two outcomes, a and b and c for example, so there would be several possible odds ratios

• the odds of a to the odds of not-a (b or c)
• the odds of b to the odds of not-b (a or c)
• the odds of c to the odds of not-c (a or b)
• the odds of a to the odds of b, ignoring c
• the odds of b to the odds of c, ignoring b
• ...

For multinomial logistic regression, the relative risk ratio is the natural generalization of the odds ratio.

I'm sure there's a text somewhere that explains this better, but I don't know it at the moment.

If your underlying question is "how do I interpret mlogit output - I know what to do with the odds ratios from logistic regression" then you may find this earlier Statalist discussion helpful.

https://www.statalist.org/forums/for...it-regressions

And that brings me to the work of Richard Williams who participated in that thread. If you are not already familiar with his papers, you may find the explanation you need, and more, in the nice overview of interpretation of multiple outcome models at https://www3.nd.edu/~rwilliam/xsoc73994/Margins05.pdf. He begins with the following reassurance.

As was the case with logit models, the parameters for an ordered logit model and other multiple outcome models can be hard to interpret. Adjusted predictions and marginal effects can again make results more understandable.

I'll also note that Margins05.pdf is preceded by Margins01.pdf ... Margins04.pdf covering other topics in the use of marginal effects to understand categorical data models.

And all five of these PDFs, and plenty more of use to someone learning about the analysis of categorical data, are linked to from https://www3.nd.edu/~rwilliam/xsoc73994/index.html - the material mentioned above is found in the section headed Interpreting results: Adjusted Predictions and Marginal effects.