Regression for nominal variables?

[Franz Dobermann]

well, i'm interested if there is some kind of procedure, similar to multiple regression to calculate results for nominal variable with more than two categories.

For example, If I want to calculate predictors (and their strenght) of voting behavior (political parties) what kind of procedure I have, with similar logic as regression analysis?

Thank you for your answer.

[roscoe]

Discriminant function might do what you want.