Logistic Regression with Categorical Dependent Variable

The regression model with categorical dependent variable is a natural generalization of the model with binary dependent variable. It is based on the use of baseline logits. For its building and for the evaluation of its quality, analogous procedures to the case of binary dependent variable are applied. When the categories of dependent variable are ordered (ordinal variable) the construction of model can be based on adjacent or cumulative logits or on proportional odds. The way of building of the model influences the meaning and the interpretation of its parameters.