Overlaying maps showing zones or polygons assigned to categorical classes is a frequent and basic process in geographical information systems (GIS). Decisions or inferences made on the basis of overlaid maps can be no more sound or valid than the information contained on component maps and on the maps that are derived by combining and permuting input maps. We possess few statistical techniques for analyzing the error inherent in categorical maps and even fewer techniques for alerting those who use GIS maps to the errors inherent in the images they see. The progress that has been made in developing tools for treating continuous distributions has not been matched with understandings of the discrete categories shown on categorical maps. Professor Chrisman's research will strengthen the theory of map error and generate new statistical procedures to describe error on overlaid categorical maps. Chrisman's work will make it possible for maps to be accompanied by information on probable error. Users will then be able to factor such imprecision into the decisions and inferences they make on the basis of maps produced by overlay procedures. As part of the process of developing such indices, Chrisman will explore the development of a new class of spatial statistics with applications beyond categorical maps. His work will be an important step toward development of a general and generally- accepted model of error in cartographic data and in cartographic display.
