Anjanette M. Chan-Tack: The Case for Spatially-Sensitive Data: How Data Structures Affect Spatial Measurement and Substantive Theory[Abstract]
Innovations in GIS and spatial statistics offer exciting opportunities to examine novel questions and to revisit established theory. Realizing this promise requires investment in spatially-sensitive data. Though convenient, widely-used administrative datasets are often spatially insensitive. They limit our ability to conceptualize and measure spatial relationships, leading to problems with ecological validity and the MAUP – with profound implications for substantive theory. I dramatize the stakes using the case of supermarket red-lining in 1970 Chicago. I compare the analytical value of a popular, spatially insensitive administrative dataset with that of a custom-built, spatially sensitive alternative. I show how the former constrains analysis to a single count measure and aspatial regression, while the latter’s point data support multiple measures and spatially-sensitive regression procedures; leading to starkly divergent results. In establishing the powerful impact that spatial measures can exert on our theoretical conclusions, I highlight the perils of relying on convenient, but insensitive datasets. Concomitantly, I demonstrate why investing in spatially sensitive data is essential for advancing sound knowledge of a broad array of historical and contemporary spatial phenomena.