MULTIVARIATE MODELLING OF CATEGORICAL DEPENDENT VARIABLES: AN INTRODUCTION TO LOGISTIC REGRESSION

Abstract

This paper is going to present the basic elements of one of the most recent techniques for analyzing the categorical binary dependent variables: the logistic regression model. By utilizing a relatively self-contained approach was tried to show the models more important dimensions with no reference to complex algebraic demonstrations. Two models illustrations were developed, in summary terms, to show the computational procedures, model contrast and selection, and also the most generalized perspectives for parameters interpretation and analysis which are implied in the modeling process.
In this way is presented to the local scientific community of health and social Sciences a technical strategy which in the next years will have a considerable development becoming in a very important analytical instrument in the toolkit of these professionals.