An application of structural equational modelling on fertility in metropolitan areas of Turkey

Abstract

Main objective of this thesis is to find out social and demographic determinants offamily formation in post-transition Bulgaria. At first, the study examines the social anddemographic developments in the country as a European country and then, as one of theBalkan countries. The study which begins with the acceptation of the reality that theattitudes and approaches to family building have changed, presents a background of thecurrent situation of family formation by mentioning prevailing social and demographiccircumstances in pre and post transition Bulgaria.In this study, data of the ?Fertility and Family Survey, Bulgaria? was used. Thissurvey was conducted by Bulgarian Academy of Sciences in December, 1997. Therespondents were the women aged 18-45. Men were not involved in the survey samplebecause of the financial restrictions. During the survey 2367 women were interviewed.The thesis includes two different types of analysis: Descriptive analyses andmultivariate analyses. Multivariate analyses cover regression and ?decision tree? analyses.It is the first demographic study which employs ?decision tree? analysis which is usuallyused by the scientists working on ?data mining system? among the studies on familyformation in our country. This method is employed in order to add new details to theresults of regression analyses, thus ?intention to have a(nother) child? which is asignificant issue regarding social and demographic researches is examined. In the firstplace, data sources of the descriptive analyses are the censuses. These analyses determinethe place of Bulgaria among the other Balkan countries, social and demographicdevelopments in pre and post transition Bulgaria; it particularly presents the changes infertility, marriage and family building. In the second place, the descriptive analysesexamine basic demographic characteristics of the women who are the origins of the dataused, their reproductive behaviors and their attitudes to family building. Multivariateanalyses examine the relationship among the women's background characteristics, fertilityintentions, and consequently, their approaches to family building.

In this thesis fertility has been considered as a phenomenon revealed as a result of a decision process. Therefore this thesis mainly focuses on the fertility-related decision process of a woman which is modeled as one?s perception affects her attitudes which in turn affects her fertility behavior. Moreover, it is hypothesized that a woman?s daily-life activities play an important role in this process. Because a woman can not be isolated from the social environment that she is living in, her decision process has an effect on others? decision processes and is also affected by them. In other words, it is assumed that the social environment a woman surrounded by has an indirect effect on her fertility-related decision process.The unit of analysis of the thesis is selected as ever-married women aged 15-49 living in metropolitan areas. Ever-married women aged 15-49 living in rural areas of Turkey are used as a control group. To carry out the analysis, Turkey Demographic and Health Survey-2003 (TNSA-2003) data set is used. In addition to descriptive analysis, multi-varied analysis are also applied by using Structural Equational Modeling (SEM) that enables the examination of casual and indirect relations and to work with variables that do not directly exist in the data set (latent variables). This method provides multivariate and multi-group analysis as well as descriptive analysis.Fertility rates for metropolitan areas are estimated and it is found that metropolitan areas have fairly low levels of fertility in contrast to urban and rural areas. Descriptive analysis addresses the basic social and demographical differences between women living in metropolitan areas and in rural areas. Moreover, it shows the differences of women?s daily-life activities, attitudes towards family planning, gender related issues and knowledge on sexually transmitted infections. SEM analyses reveals that the perception of a woman has a significant effect on her attitudes and her attitudes have a significant effect on her fertility behavior as suggested by the model. Multi-group analyze, which is performed in order control affect of the environment has revealed that an environment that surrounds a woman has an affect on her fertility-related decision process.