Combining Incompatible Spatial Data
Carol A. Gotway Crawford, (Office of Workforce and Career Development, Centers for Disease Control and Prevention), email@example.com, and
Linda J. Young, (Department of Statistics, University of Florida), firstname.lastname@example.org
Many programs and studies increasingly use existing data from many different sources (e.g., surveillance systems, health registries, governmental agencies) for analysis and inference. More often than not, the data have been collected on different geographical or spatial units, and each of these may be different from the ones of interest. There are many statistical issues associated with combining such disparate data. This presentation provides an introductory overview of several such issues, including the problems that can occur when making inferences from aggregated data, the concept of spatial support, and the importance of proper uncertainty assessment. From this perspective, the utility of many different statistical approaches to the problem of combining incompatible spatial data will be assessed. We review the current statistical methods for combining incompatible spatial data including raster and geoprocessing GIS operations, centroid smoothing techniques, regression methods, multi-level tree models, Bayesian models, and geostatistical methods. Several examples from public health will illustrate the relevant statistical issues.