Other options for multinomial logistic regression

[karina14]

Hi! I have more than 2 independent variables and they are ordinal. The dependent variable is categorical with 3 categories. I want to see the effect of these independent variables on the dependent variable. I tried multinomial logistic regression but the chi-square in the overall model test is significant, so I guess it indicates lack of fit. I didn't use linear regression because the residuals distribution isn't normal.

Are there other methods that I can use?

Thanks

[mcfanda@gmail.com]

In generalized linear models, a significant chi square does not indicate lack of fit, quite the opposite