Introducing Logistic Regression

The aim of this article is to acquaint Czech sociologists who are interested in quantitative methods of data analysis with logistic regression. The logistic regression model has been in use in statistical analyses for many years. It is the standard method for regression analysis of categorical data in many fields. Unfortunately, in the Czech Republic the method is less known, and consequently rarely used. The content of the article is divided into several parts according to the type of the explained variable. The article begins with the simplest case, in which the explained variable is dichotomous. It continues with the case, in which the explained variable is polytomous and nominal, or is treated as being nominal. The article ends with the case, in which the explained variable is ordinal. Special attention is devoted to the manner of treating categorical explaining variables, to the assessment of the adequacy of the fitted model, and to the interpretation of the coefficients of the logistic regression model. All these steps are demonstrated on real data set through the use of the SPSS and GoldMineR software packages.