Abstract:
|
A popular model for causal inference is based on potential outcomes if study units receive each of the treatments in the study. A fundamental assumption under this framework is no interference; that is, the potential outcomes of one unit are not affected by the treatment of other units. This assumption does not hold in the presence of spatial autocorrelation, where we may expect spillover or diffusion effects based on units' proximity to other units. In this talk, we extend existing methods to estimate causal effects based on spatial neighborhood structure. We specifically propose estimates of direct and spillover effects. We will present results of applying the method to the Surveillance, Epidemiology, and End Results (SEER) Program. In 2005, the Environmental Protection Agency designated several SEER counties as nonattainment for fine particulate matter mass (PM2.5) quality. We will estimate the causal effects of this designation on incidence of non-Hodgkin's lymphoma within the counties.
|