Modifiable areal unit problem

The modifiable areal unit problem is the dependence of statistical results on how spatial data are grouped into zones and on the scale of those zones. It makes conclusions sensitive to arbitrary choices of spatial boundaries.

Scale effect: Variation in statistical results caused by analyzing data at different levels of spatial aggregation. Changing zone size can shift averages, correlations, and apparent spatial patterns.
Census tract: A small geographic area used to organize population and housing statistics in a census. Tract boundaries shape neighborhood statistics and can change comparisons between places.
Stan Openshaw: A British geographer known for research on spatial analysis and the modifiable areal unit problem. His work established the problem as a central issue in quantitative geography.
Individualistic fallacy: The error of inferring group-level patterns directly from relationships observed among individuals. It reverses the ecological fallacy, while the areal-unit problem concerns sensitivity to spatial grouping.
Zoning effect: Variation in statistical results caused by drawing different boundaries around areas of the same general scale. Different boundary configurations can produce different results even when zone sizes are similar.
Geographic information system: A system for storing, analyzing, and displaying data tied to geographic locations. GIS operations create and compare the zones whose boundaries can influence results.
Geography: The study of places, spatial relationships, and interactions between people and environments. The problem emerged from geographic analysis of data summarized over areas.
Sampling bias: Systematic distortion caused when the observations included in a sample do not represent the target population. Areal-unit sensitivity can occur even with complete data, unlike bias caused by an unrepresentative sample.
Ecological fallacy: The error of inferring individual-level relationships from aggregate data. Aggregation can make area-level patterns appear to describe individuals when they do not.
Spatial epidemiology: The study of how health outcomes and their determinants vary across geographic space. Disease rates and exposure estimates can change when health data are aggregated differently.

Linked from 12 pages