Can I use a multinomial logit regression analysis using cubic polynomial of time to control time dependence?

[April Kimm]

If not, which method should I use to control time dependence? My dependent variable is categorial. Thank you!

[Clyde Schechter]

Yes, you can. The question is whether it's a good idea in your circumstance or not. Perhaps if you explained more about what you are trying to model, it might be possible to give more specific advice.

[April Kimm]

Thank you. My dependent variables are A, B, and C. They are categorical. My hypotheses are A is more likely to occur in the initial phase; B is more likely to occur in the interim phase, and C is more likely to occur in the later phase.

I am not sure if I employ multinomial logit using cubic polynomial of time is a good idea in my circumstances. or should I use cubic splines to control time dependence? How about multinomial logit regression? Any suggestions would be really appreciated.