Mapping Mortality: Evaluating Color Schemes for Choropleth Maps

Use of color for representing health data on maps raises many unanswered questions. This research addresses questions about which colors allow accurate map reading and which colors map users prefer. Through the combination of a review of previous color research and an experiment designed to test specific combinations of colors on maps, criteria were established and evaluated for selecting colors for choropleth maps of mortality data. The color-selection criteria provide pairs of hues for diverging schemes that avoid naming and colorblind confusions. We also tested sequential and spectral schemes. Our results show that color is worth the extra effort and expense it adds to map making because it permits greater accuracy in map reading. In addition, people prefer color maps over monochrome maps. Interestingly, scheme preference is affected by levels of clustering within mapped distributions. In this research, people preferred spectral and purple/green hue combinations. Contrary to our expectations, spectral schemes are effective if designed to include diverging lightness steps suited to the logical structure of mapped data. Diverging schemes produce better rate retrievals than both spectral and sequential schemes, however. In addition, diverging schemes place better emphasis on map clusters than sequential schemes. Thus map effectiveness is improved by use of diverging schemes. Our interdisciplinary research connects geographers with epidemiologists through concern about map symbolization and map reading, strengthening a significant area of collaboration. Providing guidelines that improve the design of customized color schemes will assist map makers in all disciplines in gaining insights about their data.