Uygunluk Analizinde Logaritmik Doğrusal Modellerin Kullanımı: Televizyon İzleme Eğilimleri Üzerine Bir Uygulama

Abstract

One of the methods used for analyzing categorical variables arranged in contingency tables is correspondence analysis. Correspondence analysis is named as multiple correspondence analysis for three and higher dimensional contingency tables. Correspondence analysis is a method which provides a visual approach by means of graphs and enables interpretation of categorical data easily. However, correspondence analysis does not have a significance test support for testing a model. For this reason, using it with statistical methods like logarithmic linear models contributes the correspondence analysis. In the analysis of logarithmic linear models, dual or higher degree relationship and interaction between the variables with two or more categories can be tested by means of an appropriate model. Consequently, logarithmic linear models and multiple correspondence analysis correspondence analysis can be used as complement of each other.