L'analyse de données tabulaires avec les modèles lineaires et loglineaires

Abstract

Analysis of Cross-Tabulation Data with Linear and Loglinear Models. Several models have been proposed to evaluate the strength of the relationships between observed variables, some dependent on others. In particular, when the variables are categorical, models in which the effects are considered additive (linear models), or multiplicative (loglinear/logit models) are often used. In the most complète version of these models, dependence relationships between several categorical dépendent and independent variables can be modelled. The aim of this article is to provide basic knowledge of these models to be able to apply them satisfactorily. The fundamental dependency equations, the construction of the complete models, and récent developments are presented. Further, références to programs that may be used to apply these models, and to more technical works, are provided. This présentation is based on the analysis of an example.