Presentation: Space-Time Analysis of Survey Data
Speaker: Katie Wilson, Graduate Student, UW Biostatistics
Abstract: Much of our information on health outcomes come from surveys and censuses, especially in the developing world. In this talk, we focus on several issues that can arise when analyzing this type of data and methods that can be used to address these problems. Ultimately, we are interested in understanding trends in health outcomes over time and identifying where in a country disparities may exist. We first consider analyzing survey data taken at known GPS locations in combination with census data, which are outcomes aggregated over space. We next describe how to incorporate survey data, but with missing GPS locations. Finally, in the context of child mortality, we show how to use data derived from birth histories that have varying degrees of detail in terms of when births and deaths occur in time.