TRANSFORMATION OF A CONTINUOUS TO A DISCRETE VARIABLE: AN APPLICATION OF THE ORDERED LOGIT MODEL TO EXAMINE EFFECTS OF EDUCATION ON INCOME IN TURKEY

Abstract

The aim of this study is to put forward income differences for Turkey according to the ordered logit and human capital models. For this reason, data of the Budget Survey between 2002 and 2006 by TurkStat were used. The dependent variable was the annual yearly disposable income from the main job acquired by an individual, who brought income to the household from his main job. Interpolation was used in categorizing income, a continuous variable, and curve fitting was performed with the least squares approach. In the study, the probabilities in income groups and the changes in probabilities were calculated by the ordered logit model. Later on, it was aimed to put forward the factors that determine income differences for Turkey by using the information on education, age, occupation and marital status of individuals and the social and economic information for a household through the Mincer type of human capital model.