Goodness of fit for multinomial logistic regression in Complex Survey

[Galya bigman]

Hello,


I'm wondering how I can get Goodness of fit statistics when I am running a multinomial logistic regression model while controlling for survey design.

svy: mlogit

In case I can't perform Goodness of fit text on such model how another way I can test the final model to know that it is a good fit or the best


I will appreciate your response

[ericmelse]

Hello Galya,

At the following UCLA IDRE Stata instructional webpage you will find a worked example of the user-written command fitstat, which compares fit measures for two models.
However, this probably is not applicable while controlling for survey design. So, I suppose this is not helpful in your case.