Categorical Regression Based on Optimal Scaling and An Application

Abstract

Logistic regression analysis which aims to explain the models whose dependent variable is categorical is frequently used in social science studies. In comparison with alternative techniques, exiguity of assumptions and easy interpretation of outputs make logistic regression analysis attractive. Similarly, the techniques which are known optimal scaling are used in the analysis of categorical variables. Just like logistic regression, another optimal scaling technique that aims to explain the models whose dependent variable is categorical is categorical regression analysis. Also, similar to logistic regression analysis, categorical regression has few assumptions and produce highly effective solutions. In this study the structure and properties of categorical regression analysis that based on optimal scaling are investigated. For this purpose, optimal scaling techniques are explained briefly, then information about the structure of the categorical regression analysis is described. In the application part of the study, by benefiting from the study named �Sociological Analysis of Southeast Problem� by Bilgi� and Aky�rek (2009), the relationships between the confidence level in media and various demographic variables will be investigated using categorical regression analysis. Thus, it is deduced that the categorical regression could be an alternative technique to logistic regression and models involving categorical variables were concluded with the categorical regression analysis can be made.