Providing people with safe drinking water is one of the most important health-related infrastructure programs in the world. This dissertation investigates the effects of a major water improvement program in rural China on the health of adults and children. Using panel data covering about 4,500 households from 1989 to 2006, I estimate the impact of introducing village-level access to water from water plants on various measures of health. Ordinary least squares (OLS) estimation of the impact suggests a weak positive influence of the program on people's health status, but these results may be contaminated by endogenous timing and placement of the water quality interventions across China. To address potential endogeneity problems, I use topographic characteristics of communities as instruments for program placement, as these characteristics affect the costs of the construction of water plants and pipelines into villages. My instrumental variables (IV) results show that the introduction of treated plant water into villages has had a stronger impact on the health status of both adults and children. However, the IV strategy may result into overestimation due to some omitted variables. Combining both OLS and IV estimates, I find that the illness incidence of adults decreased by 11 to 50 percent and their weight-for-height increased by 0.835 to 2.580 kg/m following the program implementation. There was also an improvement in self-reports of health. Children's weight-for-height and height itself both rose, by 0.446 to 0.754 kg/m and 0.962 to 2.489 cm respectively, as a result of the program. Using a variety of robustness checks, I show that the results are not driven by measurement errors, omitted variable bias, or attrition bias, and that the mechanism by which the program was effective was via improved water quality rather than simply via increased access to water. No obvious heterogeneous treatment effects are found across income and educational groups.