Health Services, Schools, Attitudes, and Contraceptive Use: Tests of a Theoretical Model Among Rural Nepalese.

Abstract

This dissertation uses new, highly detailed measures of dimensions of social context to advance our understanding of both attitudinal and programmatic mechanisms through which social context influences individual behavior. To illustrate, I focus on the relationship between dimensions of social context, specifically health services and schools, and contraceptive use in Nepal. Using data from the Chitwan Valley Family Study I employ multiple estimation techniques including multilevel-logistic and OLS regressions and discrete-time hazard modeling. To investigate the potential role of attitudinal mechanisms I explore how the health service and school contexts influence individuals’ attitudes about contraceptive methods, family size and composition, children, family, and non-family behaviors and how both context and these attitudes independently influence individuals’ contraceptive use. I find that women with positive attitudes about contraceptive methods and less family-oriented attitudes had higher rates of contraceptive use. Furthermore, women’s attitudes about children and family had strong effects independent from those of attitudes about contraception, indicating that these less closely connected attitudes play an integral part in women’s contraceptive use. When investigating the role of programmatic mechanisms I examine the separate effects of the provision of family planning, child, and maternal health services and, for schools, the effects of curriculum, teacher gender and training, students’ gender, and education costs. Because health service providers and schools are associated with a specific place, as new ones are built or existing ones change the services they offer, the distribution of these dimensions across physical space changes. Furthermore, because information about health services and schools is transmitted throughout communities through social channels, their potential realm of influence is also geographically disperse. Consequently, I use an expanded conceptualization of the influence of social context and estimate models with geographically weighted measures of health services and schools that incorporate all the providers and schools in the study area. My analyses using these geographically weighted measures indicate that these dimensions of health services and schools all influence contraceptive behavior. I also find evidence that women’s attitudes about children, the importance of childbearing, and roles within the family are mechanisms through which health services and schools influence contraceptive use.