SOME ASPECTS OF SPECIFIKATION OF LINEAR PROBABILITY MODEL

Abstract

In many situations dependent variable in a regression equation is not continual, but discrete choice. Modeling in such situations includes cases when dependent variable takes only two values as well as cases when the choice has to be made between few possibilities. All the models can be basically classified as linear and nonlinear. The linear probability model allows simple estimation of the parameters and interpretation of the results. Besides the model with dichotomous dependent variable, in the article is presented the model with replicated data and the model with polytomous dependent variables.

Keywords

• Linear probability model