“Sponge Maps”: Using the Concept of Value by Area Maps for Avoiding the Area Size Bias in Choropleth Maps

Abstract

AbstractA well-known problem with choropleth maps is the cognitively induced effect that larger regions are perceived as more dominant. Consequently, unsatisfactory detection rates for small areas can result, which becomes relevant when important spatial features or patterns are explored (e.g., regions with maximum values). One possible approach to avoiding the area size bias is the use of cartograms. While there is already some work on Equal Are Unit Maps, little or no research has been done on the possibility of applying the concept of Value By Area Maps to transform the reference area maps. One goal of this article is to introduce the concept of so-called “Sponge Maps”, which distort the base maps independently of thematic attributes, but depending on the need to show or emphasize certain areas of interest. The second goal of the article is to answer the overall research question whether Sponge Maps actually reduce the area size bias and improve the detectability of maximum value regions. A user study was able to verify the effectiveness of the Sponge Map approach in particular. However, it also became clear that not only the area size bias plays a role in the detection of important regions—dependencies on the absolute position (top-down bias), compactness or conspicuous shape (shape bias), the familiarity (awareness bias), the color intensity (darkness bias) and not least by the distortions in the Sponge Maps as such (distortion bias) are shown. Furthermore, special aspects of detecting minimum value regions are revealed (including the so-called inverse area size bias).