Content area
Abstract
The association between population health and sociodemographic characteristics is well documented. In this dissertation, I examine three issues that contribute to a better understanding of the pathways through which these factors are linked.
Chapter 1 addresses two opposing hypotheses that have been proposed to explain the effect of education on mortality across age: cumulative advantage and age-as-leveler. I examine whether the observed converging lifecourse pattern could be an artifact of selective mortality due to unobserved heterogeneity. Findings from a simple macrosimulation model suggest that unobserved heterogeneity exerts a substantial amount of downward bias on the estimated effect of education on mortality in old age, such that an underlying cumulative effect of education on mortality across age at the individual level could appear instead as a decreasing effect in old age.
Chapter 2 examines whether the effect of education on mortality for U.S. adults differs by gender. Discrete time logit models are used to analyze a nationally representative dataset. The results show that education has a comparable effect on mortality for men and women. No statistically significant gender difference is found in all-cause mortality, mortality by cause of death, among younger persons, and among the elderly. Analyses by marital status, however, suggest that these findings apply only to married men and women. Possible explanations for these patterns are discussed.
Chapter 3 analyzes how body weight affects health ratings across age, by sex and race, and whether the relationship can be explained by health behaviors and medical conditions. Latent growth models are employed to analyze a sample of young adults who were followed for 20 years through mid-adulthood. No significant relationship between BMI and health ratings across age is found for black adults. The effect of body weight is stronger for white men and women, for whom body weight is associated with lower starting health ratings, as well as with a faster health decline in across age. The mediating covariates explain only a small part of the BMI-SRH association.
