September 2002 A CENSORED REGRESSION MODEL FOR THE EDUCATIONAL ATTAINMENT OF BOYS AND GIRLS IN TURKEY by Mehtap Hisarciklilar* School of Economics The University of Nottingham Phone: + 44 115 9515608 Fax: + 44 115 9514159 E-mail: Lexmh@nottingham.ac.uk Abstract This study aims to analyse the reasons behind the gender-based differences in educational attainment in Turkey. It focuses on the effects of family background and household characteristics on the demand for education. Using 1988 Household Survey Data, final educational achievement regressions for boys and girls are estimated by ordered probit modelling incorporating both the random family effects and censoring in the data. The results suggest that family background characteristics, particularly parents’ education level, have a significant effect on the education of individuals and that some of these effects are different for boys and girls. JEL Classification: O15, I21 * I’m grateful to Andrew McKay and Peter Wright for their valuable suggestions and comments in writing this paper. Also, special thanks to Hakan Ercan who provided the data set when I was working on my masters thesis.