Abstract:
|
Confident class micromaps use confidence intervals involving region estimates to classify regions. The estimates are compared to a common estimate or a constant. The comparisons are structured so that the class memberships are based on the intervals being below zero, including zero, or being above zero. In the case of a single variable, this approach yields a three-class map with simple appearance. Extensions to two-way and three-way confident class map designs have also been developed. In this presentation we use the dynamic TCmap java software (Temporal Change maps) to produce confident class maps that compare county subdivision statistics to the corresponding county statistics for a chosen set of counties. TCmap reads shapefiles so can address a wider range of geospatial applications. We also take a closer look at statistical inference issues that arise in such settings.
|